Which period of growth is determinant for blood pressure?

Original article 1871

Birth weight, weight change, a
nd blood pressure duringchildhood and adolescence: a school-based multiplecohort studyArnaud Chioleroa,b, Gilles Paradisa,c,d, George Madeleinee, James A. Hanleya,Fred Paccaudb and Pascal Bovetb,e
Objective We assessed the association between birth

weight, weight change, and current blood pressure (BP)

across the entire age-span of childhood and adolescence in

large school-based cohorts in the Seychelles, an island

state in the African region.

Methods Three cohorts were analyzed: 1004 children

examined at age 5.5 and 9.1 years, 1886 children at 9.1 and

12.5, and 1575 children at 12.5 and 15.5, respectively. Birth

and 1-year anthropometric data were gathered from

medical files. The outcome was BP at age 5.5, 9.1, 12.5 or

15.5 years, respectively. Conditional linear regression

analysis was used to estimate the relative contribution

of changes in weight (expressed in z-score) during

different age periods on BP. All analyses were adjusted for

height.

Results At all ages, current BP was strongly associated with

current weight. Birth weight was not significantly associated

with current BP. Upon adjustment for current weight, the

association between birth weight and current BP tended to

become negative. Conditional linear regression analyses

indicated that changes in weight during successive age

periods since birth contributed substantially to current BP at

all ages. The strength of the association between weight

Copyright © Lippincott Williams & Wilkins. Unaut

0263-6352 � 2011 Wolters Kluwer Health | Lippincott Williams & Wilkins

change and current BP increased throughout successive

age periods.

Conclusion Weight changes during any age period since

birth have substantial impact on BP during childhood and

adolescence, with BP being more responsive to recent than

earlier weight changes. J Hypertens 29:1871–1879 Q 2011

Wolters Kluwer Health | Lippincott Williams & Wilkins.

Journal of Hypertension 2011, 29:1871–1879

Keywords: birth weight, blood pressure, children, growth, life course

Abbreviation: BP, blood pressure

aDepartment of Epidemiology, Biostatistics, and Occupational Health, McGillUniversity, Montreal, Canada, bInstitute of Social and Preventive Medicine(IUMSP), University Hospital Center and University of Lausanne, Switzerland,cMcGill University Health Center Research Institute, dPublic Health Institute ofQuebec, Montreal, Canada and eMinistry of Health, Victoria, Republic ofSeychelles

Correspondence to Dr Pascal Bovet, MD, MPH, Institute of Social and PreventiveMedicine, University of Lausanne, 17 rue du Bugnon, 1005 Lausanne,SwitzerlandTel: +41 21 314 72 72; fax: +41 21 314 73 73; e-mail: [email protected]

Received 9 February 2011 Revised 5 April 2011Accepted 13 July 2011

IntroductionBlood pressure (BP) tracks across the life course [1] and

may be determined early in life, particularly during

fetal development [2]. Alterations in the intra-uterine

environment and in fetal growth could be associated with

adaptive processes that program increased susceptibility

to hypertension [3]. This hypothesis is supported by

numerous studies which have shown an inverse associ-

ation between birth weight – used as a proxy for fetal

growth – and BP [2,4]. However, some studies found no

association or a positive association between birth weight

and BP [5,6] and doubts have been raised about the

causality of this association [7]. In addition, the inverse

association between birth weight and BP may reflect

the impact of postnatal rather than fetal growth because

the reported inverse association is magnified or appears

only upon adjustment for body size measured at the time

of BP measurement [5,8,9].

Whereas rapid postnatal growth during the first 1 or

2 years of life may be critical for the determination of

future BP [10–12], studies also suggest that changes in

body weight during later years, that is in childhood and

adolescence, are of greater importance on BP in adoles-

cents [6] or young adults [13]. Singhal and Lucas [11]

hypothesized that rapid growth – for example, an upward

centile crossing or an increase in z-score of weight – early

in life (in infancy) but also during childhood could have a

detrimental effect on cardiovascular disease risk factors.

Still, the relative contribution of body weight change at

different ages on the level of current BP during childhood

and adolescence remains unclear [4]. This is due to the

paucity of large longitudinal studies with repeated

anthropometric and BP measurements at different ages

and to related statistical analytic challenges [9,14].

Previous studies analyzed one cohort at a time with

limited age ranges at which body weight or BP were

measured [6,10–13]. Since we collected several weight

measurements between birth and school age in a large

number of children from an ongoing school-based sur-

veillance program, we assessed the association between

birth weight, weight change at different ages, and current

horized reproduction of this article is prohibited.

DOI:10.1097/HJH.0b013e32834ae396

mailto:[email protected]

http://dx.doi.org/10.1097/HJH.0b013e32834ae396

C

1872 Journal of Hypertension 2011, Vol 29 No 10

BP across the entire age-span of childhood and adoles-

cence.

MethodsThe research took place in the Republic of Seychelles, a

middle-income island state in the Indian Ocean (African

region), approximately 1000 km east of Kenya. A large

majority of the population is of African descent. In the

Seychelles, child and maternal undernutrition is uncom-

mon and a high prevalence of cardiovascular disease risk

factors (e.g. hypertension and obesity) has been docu-

mented in adults [15,16] and in children [17,18].

As part of a surveillance system, data were collected

during routine medical visits at school carried out by a

dozen trained school nurses. The surveillance system has

been routinely conducted in all schools for more than a

decade and was approved by the Ministry of Health after

technical and ethical reviews. Parents or guardians pro-

vided informed consent and children were free to parti-

cipate. The Institutional Review Boards at McGill

University approved the study protocol. Nearly 100% of

children attend school up to the fourth secondary level.

Data were collected between 1998 and 2006 in children

and adolescents attending all schools in four selected

grades: kindergarten, fourth, seventh and 10th year of

obligatory school [17,18]. Children were surveyed more

than once as they moved across the four selected grades

over time. For the present study, we analyzed three cohorts

of children who had been examined twice in kindergarten

and grade 4, grade 4 and grade 7, or grade 7 and grade 10

and for whom birth data and 1-year body weight and height

were available (Fig. 1). Some of the children that were

examined on three separate occasions (at 3–4-year inter-

vals) are included in two different cohorts.

opyright © Lippincott Williams & Wilkins. Unauth

Fig. 1

G4

Cohort G4-G71yBirth

Cohort G7-G10

Cohort K-G4

5

Age

G4

0

K

1yBirth

1yBirth

BPBP

Three cohorts of children were analyzed. In the cohort K-G4, we assessed(kindergarten), and age 9.1 (grade 4) and blood pressure (BP) at age 5.5association between weight measurements at birth, age 1, age 9.2 (grade 4we assessed the association between weight measurements at birth, age 1

School nurses followed a standard measurement pro-

cedure and were regularly trained by one of the inves-

tigators (G.M.). The equipment was checked annually

for accuracy. Equipment and methods for BP and anthro-

pometry measurements were kept identical throughout

the study period. BP was measured according to a stan-

dardized protocol with oscillometric automated devices

(Omron M5; Omron), which met the international vali-

dation protocol criteria for BP measurement [19]. BP was

measured on the right arm, in the sitting position, after a

5-min rest, and using a cuff of appropriate size according

to the arm circumference. Two readings were obtained at

a 1-min interval and averaged. Elevated BP was defined

as SBP and/or DBP equal to or above the US reference

sex, age, and height-specific 95th percentile [20].

Weight was measured using precision electronic scales

(Seca 870; Seca, Hamburg, Germany) and height was

measured with fixed stadiometers (Seca 208; Seca, Ham-

burg, Germany). Overweight and obesity were defined

using the sex and age-specific BMI criteria of the CDC,

that is, body mass index (BMI) percentile at least 85th

and at least 95th, respectively [21]. Early life anthropo-

metric data (birth weight, gestational age at birth, and

body weight at 1 year) were gathered retrospectively from

medical records [22]. Gestational age was estimated from

the date of the first day of the mother’s last menstrual

period reported in medical records. Internal sex and age-

specific z-scores (i.e. standard deviation scores) for weight

and height (or length) at each age were generated.

Statistical analysesTracking of body weight z-scores was assessed by

Pearson’s correlation coefficient between two measure-

ments from the same children [1,23].

orized reproduction of this article is prohibited.

G7 G10

(years)

G7

10 15

K: kindergartenG4: grade 4G7: grade 7G10: grade 10BP

BP

the association between weight measurements at birth, age 1, age 5.5and at age 9.1, respectively. In the cohort G4–G7, we assessed the), and age 12.5 (grade 7) and BP at age 12.5. In the cohort G7–G10,, age 12.5 (grade 7), and age 15.5 (grade 10) and BP at age 15.5.

Weight change and blood pressure in children Chiolero et al. 1873

We assessed the association between birth and current

body weight z-scores, changes in body weight z-scores,

and current BP measured at 5.5 and 9.1 years in the

kindergarten–grade 4 cohort, at 12.5 years in the grade

4–7 cohort, and at 15.5 years in the grade 7–10 cohort,

respectively (Fig. 1), using standard and conditional

linear regression analyses [8,14,24]. We fitted models

within each cohort, for each sex, and for SBP and DBP

separately. We conducted sex-specific analyses because

the pattern of growth differs between boys and girls and

because the relationship between weight z-scores and BP

could differ between boys and girls. Exploratory analyses

indeed showed interactions between current weight

z-scores and sex to be significantly associated with

SBP. However, because the differences between boys

and girls were modest (see results below), we do not

emphasize the sex difference.

Standard linear regression analysesTo assess the association between birth and current body

weight z-scores and current BP, models of standard linear

regression models were fitted, following the recommen-

dations of Lucas et al. [8]. First, two models were fitted

with birth weight and current weight body z-score con-

sidered separately, that is, current BP¼aþb1�birth

weight z-score, corresponding to the ‘early’ model in

Lucas et al.’ terminology [8]; current BP¼aþb2� cur-

current body weight z-score, corresponding to the ‘late’

model. Second, a model was fitted with both body weight

z-scores considered simultaneously, that is, current

BP¼aþb1�birth weight z-scoreþb2� current body

weight z-score, corresponding to the ‘combined’ model

[8]. Furthermore, because individuals born small for

gestational age (SGA) tend to have a greater upward

body weight postnatal centile crossing (i.e. catch-up

growth) and their BP may be more sensitive to weight

change compared to individuals with normal birth weight

[25], we fitted additional models with an interaction term

between birth and current weight z-scores, that is, current

BP¼aþb1�birth weight z-scoreþb2� current body

weight z-scoreþb3� [birth weight z-score� current

body weight z-score], corresponding to the ‘interaction’

model [8]. If the interaction term turns out to be signifi-

cantly associated with BP, this means that postnatal

weight change has a differential effect on BP in SGA

and normal birth weight individuals [8]. To account for

the potential effect of height on BP, all these models

were fitted with an adjustment for birth length or/and

current height.

Conditional linear regression analysesWe fitted conditional linear regression (also called un-

explained residual regression [24]) models to assess the

relative contribution of changes in body weight z-score

during different periods on BP level. With such model,

the effect of weight change during each period is assessed

independently from the effect of weight change during


other periods [13,24,26]. This method allows estimation

of the specific contribution of weight changes on BP at

different age periods during childhood and the identifi-

cation of sensitive periods for the effect these changes

have on BP [26].

For each age of the children after birth, we first fitted

separate linear regression models of body weight z-score at

each exam on previous weight(s) and calculated residuals

of body weight z-scores at each age. These residuals

express how many units of body weight z-score was gained

or lost above the expected z-score over the prior interval.

The following conditional regression model was then

fitted: BP¼aþb0�birth weight z-scoreþb1� residual

of one year body weight z-scoreþb2� residual of inter-

mediate body weight z-scoreþb3� residual of current

body weight z-score.

For each residual, and at each age period, the regression

coefficient can be interpreted as the effect on current BP

of change in body weight during a given age period, given

past body weight history [13]. The coefficient expresses

the difference in BP between a child who has gained one

standard deviation of weight during a given age period to

a child with an identical past weight history who did not

gain weight during this period [13]. Each coefficient for

different age periods can be interpreted independently

from the other because residuals are statistically unre-

lated to each other. To account for the potential effect of

height on BP, these models were fitted with an adjust-

ment for birth length and change in height during the

considered age periods (using residuals of 1 year, inter-

mediate, and current body height z-scores, respectively).

Adjustment for patients’ age at each exam and for birth

year was made in all models.

Analyses were carried out with Stata 10.0 (StataCorpLP,

College Station, Texas, USA).

ResultsSome 2743 children were included in this study, that is,

1004 in the kindergarten–grade 4 cohort, 1886 in grade

4–7, and 1575 in grade 7–10, respectively. Table 1 shows

the characteristics of the children in each cohort. The

average time between two exams was 3.6 years (SD 0.3)

in cohort kindergarten–grade 4, 3.4 years (0.3) in cohort

grade 4–7, and 3.0 years (0.3) in cohort grade 7–10, respec-

tively.

Tracking of body weight z-score was relatively weak

between weight at birth or weight at 1 year of age and

weight later in life (Table 2). Tracking was much stronger

between successive body weight measurements in child-

hood and adolescence.

Using standard linear regression analyses, neither SBP

(Table 3) nor DBP (Table 4) was associated with birth

weight z-score. In each cohort, a strong association

was found between current body weight and BP. The


C


Table 1 Characteristics of the participants in the three cohorts

Cohort K–G4 (N¼1004) Cohort G4–G7 (N¼1886) Cohort G7–G10 (N¼1575)

K G4 G4 G7 G7 G10

BoysN 522 930 738Birth weight (kg) 3.1 (0.5) 3.2 (0.5) 3.2 (0.5)Gestational age (week) 39.2 (1.6) 39.4 (1.5) 39.4 (1.5)One year weight (kg) 10.0 (1.2) 10.0 (1.1) 10.0 (1.1)Age (year) 5.5 (0.4) 9.1 (0.4) 9.1 (0.4) 12.5 (0.4) 12.5 (0.4) 15.5 (0.4)Weight (kg) 19.5 (3.4) 29.6 (7.3) 29.0 (6.2) 42.7 (10.7) 41.6 (10.1) 57.7 (11.2)Height (cm) 113.4 (5.6) 134.0 (6.5) 134.0 (6.3) 152.3 (8.3) 151.2 (8.1) 169.8 (7.3)BMI (kg/m2) 15.1 (1.9) 16.4 (3.1) 16.1 (2.7) 18.3 (3.6) 18.1 (3.4) 20.0 (3.4)Overweight (%) 10.3 14.7 11.4 16.1 14.1 9.9Obesity (%) 5.6 7.9 5.7 6.5 5.7 4.1SBP (mmHg) 95.1 (10.0) 100.6 (11.2) 101.0 (11.0) 104.0 (11.3) 104.4 (10.8) 115.6 (12.2)DBP (mmHg) 58.4 (9.5) 62.3 (9.8) 61.9 (9.7) 64.0 (8.4) 64.8 (8.8) 68.5 (8.6)Elevated BP (%) 6.7 9.0 8.5 4.8 5.0 10.3

GirlsN 482 956 837Birth weight (kg) 3.1 (0.5) 3.1 (0.5) 3.1 (0.5)Gestational age (week) 39.3 (1.6) 39.4 (1.5) 39.2 (1.6)One year weight (kg) 9.5 (1.2) 9.5 (1.2) 9.4 (1.2)Age (year) 5.5 (0.3) 9.1 (0.3) 9.1 (0.3) 12.5 (0.4) 12.5 (0.4) 15.5 (0.4)Weight (kg) 19.2 (3.4) 30.2 (7.3) 29.8 (7.1) 47.0 (11.6) 45.7 (11.4) 54.9 (11.4)Height (cm) 112.7 (5.3) 134.3 (6.6) 134.1 (6.5) 155.0 (7.1) 153.8 (7.1) 161.1 (6.1)BMI (kg/m2) 15.1 (2.0) 16.6 (3.3) 16.4 (3.1) 19.5 (4.2) 19.2 (4.1) 21.2 (4.2)Overweight (%) 10.4 16.4 14.6 21.5 19.6 18.3Obesity (%) 4.8 7.7 5.9 9.1 7.9 5.7SBP (mmHg) 94.9 (10.7) 101.2 (11.6) 101.1 (10.9) 105.8 (11.0) 106.1 (10.6) 109.1 (10.5)DBP (mmHg) 58.6 (9.6) 63.7 (9.5) 63.0 (9.4) 66.9 (8.6) 67.1 (8.8) 69.5 (7.8)Elevated BP (%) 9.5 11.2 11.0 8.8 9.1 7.9

Values are mean (SD) or proportion. BP, blood pressure; G, grade; K, kindergarten. Overweight and obesity were defined using the sex- and age-specific BMI criteria of theCDC, that is, BMI percentile at least 85th and at least 95th, respectively [21]. Elevated BP was defined as SBP and/or DBP equal to or above the US reference sex, age, andheight-specific 95th percentile [20].

association between current weight z-score and SBP was

slightly weaker at younger (5.5 or 9.1 years) than at later

ages (12.5 or 15.5 years) (Table 3). The association

between current weight z-score and BP was generally

weaker for DBP than for SBP. For DBP, the strength of

the association with current body weight z-score was

similar at all ages (Table 4). These patterns were similar

in both sexes.

In standard linear regression analyses considering simul-

taneously birth and current body weight measurements,

the association between birth weight z-score and BP

tended to decrease and, in some cases, become negative.

In these models, current weight remained strongly associ-

ated with BP. In each cohort, the interaction term

between birth weight z-score and current weight z-score


Table 2 Tracking correlation coefficient of body weight z-scorebetween successive visits

Boys Girls

Age(years) 1 5.5 9.1 12.5 15.5 1 5.5 9.1 12.5 15.5

0 0.28 0.19 0.15 0.12 0.09 0.27 0.22 0.18 0.15 0.181 0.52 0.42 0.33 0.29 0.60 0.42 0.39 0.375.5 0.84 0.77 0.81 0.739.1 0.87 0.76 0.87 0.7812.5 0.81 0.84

was not statistically significantly associated with SBP or

DBP (P-values for interaction term 0.11–0.98).

Conditional regression analyses showed that both SBP

and DBP were positively associated with body weight z-

score change in all the considered age periods after birth

(Figs 2 and 3, Online supplement Table S1 and S2, http://

links.lww.com/HJH/A113) but the strength of the associ-

ation between weight gain and current BP tended to

increase throughout successive age periods in all cohorts.

The strength of the association was generally weaker for

weight change between birth and 1 year of age than for

later weight change. These findings were similar for boys

and girls and for SBP (Fig. 2) and DBP (Fig. 3), albeit the

strength of the associations was generally weaker for DBP

than SBP.

DiscussionThe major finding of our study is that weight changes

during successive age periods since birth contributed

substantially to current BP during childhood and adoles-

cence and that this association was stronger for weight

changes in recent than earlier years of children’s growth.

Weight change in the first year of life had a significant but

relatively small impact on current BP. No significant

association was observed between birth weight and cur-

rent BP although there was a trend toward negative


http://links.lww.com/HJH/A113

http://links.lww.com/HJH/A113

Copyright © Lippincott Williams & Wilkins. Unauthorized reproduction of this article is prohibited.


Ta

ble

3R

eg

ress

ion

coe

ffici

en

t(b

)e

stim

ate

sfr

om

reg

ress

ion

ofc

urr

en

tSB

Po

nb

irth

or

curr

en

tbo

dy

we

igh

tz-s

core

con

sid

ere

dse

pa

rate

ly(I

)a

nd

fro

mre

gre

ssio

no

fcu

rre

ntS

BP

on

bir

tha

nd

curr

en

tb

od

yw

eig

ht

z-sc

ore

sco

nsi

de

red

sim

ult

an

eo

usl

y(I

I)

Bo

ysG

irls

(I)

Dis

tinct

mo

del

sw

ithb

irth

or

curr

ent

wei

ght

(II)

Mo

del

sw

ithb

irth

and

curr

ent

wei

ght

(I)

Dis

tinct

mo

del

sw

ithb

irth

or

curr

ent

wei

ght

(II)

Mo

del

sw

ithb

irth

and

curr

ent

wei

ght

Co

hort

Ag

eat

BP

mea

sure

men

t[y

ears

,m

ean

(SD

)]W

eig

htz-

sco

re[m

mH

g,b

(95

%C

I)]

P[m

mH

g,b

(95

%C

I)]

P[m

mH

g,b

(95

%C

I)]

P[m

mH

g,b

(95

%C

I)]

P

K–

G4

5.5

(0.4

)A

tb

irth

0.9

(�0

.1to

1.9

)0

.07

10

.6(�

0.4

to1

.5)

0.2

71

�0

.3(�

1.4

to0

.8)

0.6

27

�0

.5(�

1.7

to0

.6)

0.3

59

Cur

rent

1.1

(0.1

to2

.2)

0.0

38

1.1

(0.0

to2

.1)

0.0

52

1.4

(0.3

to2

.6)

0.0

17

1.5

(0.3

to2

.7)

0.0

12

K–

G4

9.1

(0.4

)A

tb

irth

0.5

(�0

.6to

1.6

)0

.37

30

.0(�

1.1

to1

.1)

0.9

53

�0

.3(�

1.5

to1

.0)

0.6

70

�0

.9(�

2.0

to0

.2)

0.1

17

Cur

rent

2.8

(1.7

to3

.9)

0.0

00

2.8

(1.7

to3

.9)

0.0

00

4.5

(3.3

to5

.7)

0.0

00

4.6

(3.4

to5

.8)

0.0

00

G4

–G

71

2.5

(0.4

)A

tb

irth

0.3

(�0

.5to

1.2

)0

.43

6�

0.5

(�1

.3to

0.2

)0

.16

1�

0.3

(�1

.1to

0.5

)0

.45

2�

0.9

(�1

.6to�

0.1

)0

.02

6C

urre

nt3

.9(2

.7to

5.1

)0

.00

04

.1(3

.3to

5.0

)0

.00

02

.4(1

.1to

3.7

)0

.00

03

.4(2

.6to

4.2

)0

.00

0G

7–

G1

01

5.5

(0.4

)A

tb

irth

0.4

(�0

.9to

1.8

)0

.52

1�

0.1

(�1

.4to

1.1

)0

.84

80

.5(�

0.6

to1

.6)

0.3

75

0.1

(�1

.0to

1.2

)0

.86

8C

urre

nt4

.3(3

.1to

5.6

)0

.00

04

.3(3

.0to

5.5

)0

.00

02

.5(1

.5to

3.5

)0

.00

02

.5(1

.6to

3.5

)0

.00

0

G,

gra

de;

K,

kind

erg

arte

n.

Ta

ble

4R

eg

ress

ion

coe

ffici

en

t(b

)e

stim

ate

sfr

om

reg

ress

ion

ofc

urr

en

tDB

Po

nb

irth

or

curr

en

tbo

dy

we

igh

tz-s

core

con

sid

ere

dse

pa

rate

ly(I

)a

nd

fro

mre

gre

ssio

no

fcu

rre

ntD

BP

on

bir

tha

nd

curr

en

tb

od

yw

eig

ht

z-sc

ore

sco

nsi

de

red

sim

ult

an

eo

usl

y(I

I)

Bo

ysG

irls

(I)

Dis

tinct

mo

del

sw

ithb

irth

or

curr

ent

wei

ght

(II)

Mo

del

sw

ithb

irth

and

curr

ent

wei

ght

(I)

Dis

tinct

mo

del

sw

ithb

irth

or

curr

ent

wei

ght

(II)

Mo

del

sw

ithb

irth

and

curr

ent

wei

ght

Co

hort

Ag

eat

BP

mea

sure

men

t[y

ears

,m

ean

(SD

)]W

eig

htz-

sco

re[m

mH

g,b

(95

%C

I)]

P[m

mH

g,b

(95

%C

I)]

P[m

mH

g,b

(95

%C

I)]

P[m

mH

g,b

(95

%C

I)]

P

K–

G4

5.5

(0.4

)A

tb

irth

0.2

(�0

.8to

1.1

)0

.74

9�

0.1

(�1

.1to

0.8

)0

.76

80

.0(�

1.0

to1

.1)

0.9

33

�0

.2(�

1.2

to0

.9)

0.7

65

Cur

rent

1.0

(0.0

to2

.0)

0.0

58

1.0

(0.0

to2

.0)

0.0

55

1.5

(0.5

to2

.6)

0.0

04

1.6

(0.5

to2

.6)

0.0

04

K–

G4

9.1

(0.4

)A

tb

irth

0.3

(�0

.7to

1.3

)0

.51

2�

0.2

(�1

.1to

0.8

)0

.70

8�

0.4

(�1

.4to

0.6

)0

.38

5�

0.9

(�1

.8to

0.1

)0

.06

8C

urre

nt2

.3(1

.4to

3.2

)0

.00

02

.3(1

.4to

3.2

)0

.00

03

.1(2

.2to

4.1

)0

.00

03

.2(2

.2to

4.2

)0

.00

0G

4–

G7

12

.5(0

.4)

At

birt

h0

.0(�

0.6

to0

.7)

0.8

95

�0

.4(�

1.0

to0

.2)

0.2

03

�0

.2(�

0.9

to0

.4)

0.4

97

�0

.7(�

1.3

to�

0.1

)0

.02

8C

urre

nt2

.8(1

.9to

3.8

)0

.00

03

.0(2

.3to

3.7

)0

.00

03

.0(2

.1to

3.9

)0

.00

03

.4(2

.8to

4.0

)0

.00

0G

7–

G1

01

5.5

(0.4

)A

tb

irth

�0

.2(�

1.1

to0

.8)

0.7

24

�0

.4(�

1.4

to0

.5)

0.3

61

�0

.1(�

0.9

to0

.7)

0.7

91

�0

.5(�

1.3

to0

.3)

0.2

37

Cur

rent

1.8

(0.9

to2

.8)

0.0

00

1.9

(0.9

to2

.8)

0.0

00

2.1

(1.4

to2

.8)

0.0

00

2.2

(1.4

to2

.9)

0.0

00

G,

gra

de;

K,

kind

erg

arte

n.

C


Fig. 2

-2-101234567

Con

ditio

nal r

egre

ssio

n co

effi

cein

t(m

mH

g/un

it z-

scor

e)

Boys

Girls

at 12.5 years

Age period (year)

Systolic BP at 9.1 yearsat 5.5 years at 15.5 years

Regression coefficient estimates (and 95% confidence interval) from regression of current SBP on each specific residual of body weight z-score.

association upon adjustment for current weight. There

was no evidence that the effect of postnatal weight

change on current BP differed between individuals born

small for gestational age and individuals born with normal

birth weight.

Few studies have assessed the association between cur-

rent BP and growth taking place after infancy or after

early childhood [4,6,13]. For example, using standard

multiple regression analysis, Horta et al. [6] reported that

rapid weight gain before and after the age of 4 was

associated with high SBP at 15 years of age. Only weight

gain after the age of 4 years was associated with DBP. In

young adults, using various statistical methods, a change

in BMI z-scores after – but not before – the age of

11–13 years was shown to be a strong determinant of BP


Fig. 3

−2

−1

0

1

2

3

4

5

6

7

Co

nd

itio

nal

reg

ress

ion

co

effi

cein

t(m

mH

g/u

nit

z-s

core

)

aDiastolic BP at 9.1 yearsat 5.5 years

Regression coefficient estimates (and 95% confidence interval) from regres

at 30–40 years of age [13]. Limitations of these studies

include the relatively long interval of time between

weight measurements before BP determination and

the relatively small sample sizes. These studies may

indicate critical periods with respect of the effect of

growth on BP, but they also suggest that weight gain

that immediately precedes BP measurement is most

strongly associated with BP, that is, recent weight gain

has more effect on BP than earlier weight change, which

is consistent with our findings.

Although the effect was relatively small in our study,

weight gain during the first year of life contributed to BP

independently of later weight gain. This may be due to

tracking of body weight [23] as well as ‘programming’

of BP early in life. Whereas experimental studies in


Boys

Girls

t 12.5 years

Age period(year)

at 15.5 years

sion of current DBP on each specific residual of body weight z-score.


newborns have suggested that BP may be ‘programmed’

by faster weight gain in the first months of life [11,27],

observational studies assessing the effect of growth early

in postnatal life on later BP have provided contrasting

results [10,12,28–30]. Some studies suggest that weight

gain in the first months of life is associated with elevated

BP in young adults [12,30]. For example, it was reported

that rapid weight gain during infancy (0–6 months) – but

not during early childhood (3–6 years) – predicted risk

factors clustering associated with the metabolic syn-

drome, including elevated BP, in adolescents aged

17 years [31]. However, other studies did not find that

early weight gain was critical for later BP [10,28,29]. For

example, Law et al. [10] found that weight gain during

infancy (0–1 year) was not associated with SBP at 22 years

of age although they did find an association with weight

gain during early childhood (1–5 years). Similarly, in

children aged 3 in Britain [28] and in children aged 6

in Australia [29], BP was not associated with weight gain

during the first year of life once birth weight or current

body weight was taken into account. Overall, it remains

difficult to ascertain with confidence whether weight

gain in the first years of life is critical for later BP or

not, and our study suggests that it may be less critical than

weight gain later in life. Long-term follow-up of partici-

pants randomized to interventions potentially affecting

early growth and excessive weight gain (e.g. breastfeed-

ing vs. formula feeding) may shed light on this issue

[8,11].

We found that the association between birth weight and

current BP was null or only modest without adjustment

for current body weight and tended to become negative

upon adjustment for current weight. With regards to the

unadjusted association, other studies reported mixed

results, that is, positive, negative or no association

[5,6,32]. In a systematic review pooling data from more

than 350 000 children and adults, the (unadjusted) associ-

ation between birth weight and later BP was weakly

negative [5]. Our finding of a tendency toward a negative

association between birth weight and current BP upon

adjustment for current body weight, is consistent with

most studies showing that, after adjustment for current

body size, a negative association between birth weight

and later BP emerges or, if already present in the unad-

justed analysis, is magnified [5,9,32]. This suggests that

among children who attained the same weight at a given

age, those who were of lower birth weight have, on

average, higher BP, that is, that growth trajectories since

from conception to childhood or adolescence may be

determinants for the level of BP [33].

Blood pressure seems to be increasingly responsive to

weight changes throughout successive age periods during

childhood and adolescence. This may be due to the effect

of weight changes on attained weight and corresponding

amount of adiposity. Biologically, a large weight gain

during childhood reflects in part an accumulation of


adipose tissue. Excess adiposity is associated with inflam-

mation, production of adipocytokines, sympathetic

hyperactivity, and insulin resistance, which all may

increase BP in children and adolescents [34]. In adults,

studies suggest that BP is more sensitive to recent than

past weight decrease, and this effect has been linked to

water and salt losses and, perhaps, reduced sympathetic

activity and insulin resistance [35]. Changes in diet or

in physical activity that are concomitant with recent

body weight changes may also influence BP. Postnatal

‘programming’ of BP may also be involved. It implies the

modification or impaired development of permanent

somatic structures or the resetting of physiological sys-

tems by a given stimulus at a sensitive time period,

resulting in long-term consequences on BP [3]. For

instance, the kidneys play a major role in the long-term

regulation of BP and biological mechanisms for a

hypothetical postnatal programming of BP could involve

alterations in renal function [36]. Thus, excess weight

gain during successive age period may be associated

with the progressive development of subtle renal

deterioration due to adiposity-related factor such as

an increased intrarenal pressure or hyperuricemia [36].

Progressive increase in arterial stiffness could also be

implicated [37] and could explain that the strength of

the association between current body weight and BP

increased with age.

Our study has limitations. Our analyses are conditioned

on the age at which data were collected and we could

analyze a limited numbers of defined periods of growth.

Birth weight, gestational age and body weight at the age

of 1 year were gathered retrospectively. We did not take

into account several potential explanatory or confounding

factors (e.g. physical activity, diet, pubertal development/

sexual maturation, socio-economic factors, or parents’

characteristics), which limits causal inferences. We did

not have data on metabolic or cardiovascular character-

istics which could have provided further insight on the

cardiovascular consequences of weight change during

childhood and adolescence. Detailed information on

body composition and fat distribution would have also

helped understand biological mechanisms underlying the

association between weight change and BP [38]. Findings

in one population may not directly apply to all popu-

lations since the ideal growth trajectory for cardiovascular

health may differ among different populations [32]. Our

study also has unique strengths. We studied three large

population-based cohorts and used identical and standar-

dized measurement methods allowing a comprehensive

study of growth and BP across the entire age range of

childhood and adolescence. Our statistical methods allow

the direct comparison of the effect on current BP of

weight change during successive age periods.

Our findings have important clinical and public health

significance for the prevention and control of elevated

BP. On one hand, because BP is more responsive to


C


recent than earlier weight change, prevention of exces-

sive weight gain even relatively late in childhood

or adolescence may be important strategies to address

elevated BP or prevent its occurrence. On the contrary,

because weight gain at all ages is associated with

both increased weight and increased BP later in life,

maintenance of a healthy weight at all ages will

help prevent hypertension in childhood or adolescence

[20]. Nonetheless, it remains unclear whether interven-

tions to prevent weight gain are more efficient at younger

than at later age [39]. Further studies are required to

determine modalities and best timing for weight control

interventions during childhood as a mean to prevent

elevated BP.

AcknowledgmentsData for this study were collected within the routine

operation of an ongoing School Health Program con-

ducted by the Ministry of Health of the Republic of

Seychelles. Equipment for this program was partially

funded by the World Health Organization and some

other nonprofit organizations. We thank all the nurses

and other health officers involved with the School Health

Program and we are grateful to the Ministry of Health for

continued support to epidemiological research related to

non communicable diseases. Arnaud Chiolero is sup-

ported by the Canadian Institutes of Health Research

(CIHR). Gilles Paradis holds a CIHR Applied Public

Health Research Chair. The views expressed by the

authors of this paper do not reflect necessarily those of

their institutions.

Conflicts of interestThere are no conflicts of interest.

References1 Chen X, Wang Y. Tracking of blood pressure from childhood to adulthood:

a systematic review and meta-regression analysis. Circulation 2008;117:3171–3180.

2 Barker DJP, Bull AR, Osmmond C, Simmonds SJ. Fetal and placental sizeand risk of hypertension later in life. BMJ 1990; 301:259–262.

3 McMillen IC, Robinson JS. Developmental origins of the metabolicsyndrome: prediction, plasticity, and programming. Physiol Rev 2005;85:571–633.

4 Lawlor DA, Smith GD. Early life determinants of adult blood pressure. CurrOpin Nephrol Hypertens 2005; 14:259–264.

5 Huxley R, Neil A, Collins R. Unravelling the fetal origins hypothesis: is therereally an inverse association between birthweight and subsequent bloodpressure? Lancet 2002; 360:659–665.

6 Horta BL, Barros FC, Victora CG, Cole TJ. Early and late growth and bloodpressure in adolescence. J Epidemiol Community Health 2003; 57:226–230.

7 Kramer MS. Invited commentary: association between restricted fetalgrowth and adult chronic disease: is it causal? Is it important? Am JEpidemiol 2000; 152:605–608.

8 Lucas A, Fewtrell MS, Cole TJ. Fetal origins of adult disease-the hypothesisrevisited. BMJ 1999; 319:245–249.

9 Tu YK, West R, Ellison GT, Gilthorpe MS. Why evidence for the fetal originsof adult disease might be a statistical artifact: the ‘reversal paradox’ for therelation between birth weight and blood pressure in later life. Am JEpidemiol 2005; 161:27–32.

10 Law CM, Shiell AW, Newsome CA, Syddall HE, Shinebourne EA, FayersPM, et al. Fetal, infant, and childhood growth and adult blood pressure: alongitudinal study from birth to 22 years of age. Circulation 2002;105:1088–1092.


11 Singhal A, Lucas A. Early origins of cardiovascular disease: is there aunifying hypothesis? Lancet 2004; 363:1642–1645.

12 Ben-Shlomo Y, McCarthy A, Hughes R, Tilling K, Davies D, Smith GD.Immediate postnatal growth is associated with blood pressure in youngadulthood: theBarryCaerphilly GrowthStudy. Hypertension2008;52:638–644.

13 Gamborg M, Andersen PK, Baker JL, Budtz-Jørgensen E, Jørgensen T,Jensen G, Sørensen TI. Life course path analysis of birth weight, childhoodgrowth, and adult systolic blood pressure. Am J Epidemiol 2009;169:1167–1178.

14 De Stavola BL, Nitsch D, dos Santos Silva I, McCormack V, Hardy R, MannV, et al. Statistical issues in life course epidemiology. Am J Epidemiol 2006;163:84–96.

15 Danon-Hersch N, Chiolero A, Shamlaye C, Paccaud F, Bovet P.Decreasing association between body mass index and blood pressure overtime. Epidemiology 2007; 18:493–500.

16 Bovet P, Romain S, Shamlaye C, Mendis S, Darioli R, Riesen W, et al.Divergent fifteen-year trends in traditional and cardiometabolic risk factorsof cardiovascular diseases in the Seychelles. Cardiovasc Diabetol 2009;8:34.

17 Chiolero A, Madeleine G, Gabriel A, Burnier M, Paccaud F, Bovet P.Prevalence of elevated blood pressure and association with overweight inchildren of a rapidly developing country. J Hum Hypertens 2007; 21:120–127.

18 Chiolero A, Paradis G, Madeleine G, Hanley JA, Paccaud F, Bovet P.Discordant secular trends in elevated blood pressure and obesity inchildren and adolescents in a rapidly developing country. Circulation 2009;119:558–565.

19 El Assaad MA, Topouchian JA, Asmar RG. Evaluation of two devices forself-measurement of blood pressure according to the internationalprotocol: the Omron M5-I and the Omron 705IT. Blood Press Monit 2003;8:127–133.

20 National High BP Education Program Working Group on High BloodPressure in Children and Adolescents. The fourth report on the diagnosis,evaluation, and treatment of high blood pressure in children andadolescents. Pediatrics 2004; 114 (2 Suppl 4th Report):S555–S576.

21 Kuczmarski RJ, Ogden CL, Grummer-Strawn LM, Flegal KM, Guo SS,Wei R, et al. CDC growth charts: United States. Adv Data 2000; 314:1–27.

22 Stettler N, Bovet P, Shamlaye H, Zemel BS, Stallings VA, Paccaud F.Prevalence and risk factors for overweight and obesity in children fromSeychelles, a country in rapid transition: the importance of early growth. IntJ Obes Relat Metab Disord 2002; 26:214–219.

23 Singh AS, Mulder C, Twisk JW, van Mechelen W, Chinapaw MJ. Trackingof childhood overweight into adulthood: a systematic review of theliterature. Obes Rev 2008; 9:474–488.

24 Keijzer-Veen MG, Euser AM, van Montfoort N, Dekker FW, VandenbrouckeJP, Van Houwelingen HC. A regression model with unexplained residualswas preferred in the analysis of the fetal origins of adult diseaseshypothesis. J Clin Epidemiol 2005; 58:1320–1324.

25 Adair LS, Cole TJ. Rapid child growth raises blood pressure in adolescentboys who were thin at birth. Hypertension 2003; 41:451–456.

26 Adair LS, Martorell R, Stein AD, Hallal PC, Sachdev HS, Prabhakaran D,et al. Size at birth, weight gain in infancy and childhood, and adult bloodpressure in 5 low- and middle-income-country cohorts: when does weightgain matter? Am J Clin Nutr 2009; 89:1383–1392.

27 Singhal A, Cole TJ, Lucas A. Early nutrition in preterm infants and later bloodpressure: two cohorts after randomised trials. Lancet 2001; 357:413–419.

28 Whincup PH, Bredow M, Payne F, Sadler S, Golding J. Size at birth andblood pressure at 3 years of age. The Avon Longitudinal Study ofPregnancy and Childhood (ALSPAC). Am J Epidemiol 1999; 149:730–739.

29 Blake KV, Gurrin LC, Evans SF, Beilin LJ, Stanley FJ, Landau LI, NewnhamJP. Adjustment for current weight and the relationship between birthweight and blood pressure in childhood. J Hypertens 2000; 18:1007–1012.

30 Jarvelin MR, Sovio U, King V, Lauren L, Xu B, McCarthy MI, et al. Early lifefactors and blood pressure at age 31 years in the 1966 northern Finlandbirth cohort. Hypertension 2004; 44:838–846.

31 Ekelund U, Ong KK, Linne Y, Neovius M, Brage S, Dunger DB, et al.Association of weight gain in infancy and early childhood withmetabolic risk in young adults. J Clin Endocrinol Metab 2007;92:98–103.

32 Victora CG, Adair L, Fall C, Hallal PC, Martorell R, Richter L, et al. Maternaland child undernutrition: consequences for adult health and human capital.Lancet 2008; 371:340–357.



33 Lawlor DA, Leon DA, Rasmussen F. Growth trajectory matters: interpretingthe associations among birth weight, concurrent body size, and systolicblood pressure in a cohort study of 378,707 Swedish men. Am J Epidemiol2007; 165:1405–1412.

34 Lambert M, O’Loughlin J, Delvin EE, Levy E, Chiolero A, Paradis G.Association between insulin, leptin, adiponectin and blood pressure inyouth. J Hypertens 2009; 27:1025–1032.

35 Kuller LH. Weight loss and reduction of blood pressure and hypertension.Hypertension 2009; 54:700–701.

36 Johnson RJ, Rodriguez-Iturbe B, Kang DH, Feig DI, Herrera-Acosta J. Aunifying pathway for essential hypertension. Am J Hypertens 2005;18:431–440.


37 Sakuragi S, Abhayaratna K, Gravenmaker KJ, O’Reilly C, Srikusalanukul W,Budge MM, et al. Influence of adiposity and physical activity on arterialstiffness in healthy children: the lifestyle of our kids study. Hypertension2009; 53:611–616.

38 Lawlor DA, Benfield L, Logue J, Tilling K, Howe LD, Fraser A,et al. Association between general and central adiposity inchildhood, and change in these, with cardiovascular riskfactors in adolescence: prospective cohort study. BMJ 2010;341:c6224.

39 Gidding SS. Assembling evidence to justify prevention of atherosclerosisbeginning in youth. Circulation 2010; 122:2493–2494.


Which period of growth is determinant for blood pressure?

Documents

Transcript of Which period of growth is determinant for blood pressure?