Content uploaded by Pamela Gaskin
Author content
All content in this area was uploaded by Pamela Gaskin on May 28, 2014
Content may be subject to copyright.
Content uploaded by Pamela Gaskin
Author content
All content in this area was uploaded by Pamela Gaskin
Content may be subject to copyright.
The eVects of birth weight and postnatal linear
growth retardation on blood pressure at age 11–12
years
S P Walker, P Gaskin, C A Powell, F I Bennett, T E Forrester, S Grantham-McGregor
Abstract
Study objective—To determine the eVects
of birth weight and linear growth retarda-
tion (stunting) in early childhood on blood
pressure at age 11–12 years.
Design—Prospective cohort study.
Setting—Kingston, Jamaica.
Participants—112 stunted children
(height for age < −2 SD of the NCHS
references) and 189 non-stunted children
(height for age > −1 SD), identified at age
9–24 months by a survey of poor neigh-
bourhoods in Kingston.
Main results—Current weight was the
strongest predictor of systolic blood pres-
sure (â= 4.90 mm Hg/SD weight 95%CI
3.97, 5.83). Birth weight predicted systolic
blood pressure (â= −1.28 mm Hg/SD
change in birth weight, 95% CI −2.17,
−0.38) after adjustment for current
weight. There was a significant negative
interaction between stunting in early
childhood and current weight indicating a
larger eVect of increased current weight in
children who experienced linear growth
retardation in early childhood. There was
no interaction between birth weight and
current weight. The increase in blood
pressure from age 7 to age 11–12 was
greater in children with higher weight at
age 11–12 and less in children with higher
birth weight and weight at age 7.
Conclusions—Birth weight predicted
systolic blood pressure in Jamaican chil-
dren aged 11–12. Postnatal growth retar-
dation may potentiate the relation
between current weight and blood pres-
sure. Greater weight gain between ages 7
and 11 was associated with a greater
increase in systolic blood pressure. The
relation between growth and later blood
pressure is complex and has prenatal and
postnatal components.
(J Epidemiol Community Health 2001;55:394–398)
Barker and colleagues have suggested that
there may be developmental periods in utero
and in infancy during which inappropriate
nutrition and poor growth lead to long term
consequences for adult cardiovascular dis-
ease.12 Blood pressure in adults is typically
inversely associated with birth weight.13 This
inverse relation has also been described in
childhood once adjustment is made for current
size,4although it has not been found in some
populations.5In two previous Jamaican studies
blood pressure in childhood was related to
maternal nutritional status6and size at birth.7
Growth during infancy and early childhood
may also influence later blood pressure. In a
few studies, weight in infancy was not inde-
pendently related to later blood pressure.8–10
These studies, conducted in the UK and
France, would not have included many infants
whose postnatal growth retardation was com-
parable to that experienced by many children
in developing countries.
Many developing countries are experiencing
an epidemiological transition from a pattern of
primarily infectious diseases to increasing inci-
dence of adult chronic diseases.11 It is possible
that under-nutrition in childhood combined
with better conditions in adulthood may
contribute to the increasing levels of cardiovas-
cular disease in these countries.12
In 1986 we began a longitudinal study of 127
children aged 9–24 months whose linear
growth was retarded (stunted group, height for
age < −2 SD of the NCHS references13), and a
comparison group of children from the same
poor communities in Kingston, Jamaica whose
linear growth was normal (non-stunted group,
height for age > −1 SD). In the first two years
of the study we evaluated the eVects of supple-
mentation and psychosocial stimulation on the
stunted children’s growth and mental develop-
ment.14 15 The children were re-measured four
years after the interventions ended when they
were 7–8 years old. The height for age of the
stunted children had increased from a mean of
−3.0 SD scores on enrollment to −1.1 SD
scores at follow up.16 This suggests that, after
the period of early growth retardation, their
growth was not limited by nutritional intakes.
This cohort of children who experienced
early growth retardation followed by adequate
growth provides an opportunity to further our
understanding of the importance of growth,
both in utero and postnatally, to later blood
pressure. The objectives of this study were to
determine whether at age 11–12 years the
blood pressures of children who were stunted
diVered from those of children who were never
stunted, and to determine the relations of birth
weight and postnatal growth with blood
pressure at age 11–12.
Methods
PARTICIPANTS
The children were identified at age 9–24
months by house to house survey of several
poor Kingston neighbourhoods. At that time,
127 stunted children (height for age < −2 SD
J Epidemiol Community Health 2001;55:394–398394
Epidemiology
Research Unit,
Tropical Medicine
Research Institute,
University of the West
Indies, Mona, Kingston
7, Jamaica
S P Walker
P Gaskin
C Powell
Tropical Metabolism
Research Unit,
Tropical Medicine
Research Institute,
University of the West
Indies
F I Bennett
T E Forrester
Institute of Child
Health, London, UK
S Grantham-McGregor
Correspondence to:
Professor Walker
(swalker@uwimona.edu.jm)
Accepted for publication
9 February 2001
www.jech.com
of the NCHS references) and 32 non-stunted
children (height for age > −1 SD) participated
in a two year intervention study. The stunted
group comprised all stunted children identified
and the non-stunted children were matched to
every fourth stunted child. Four years later,
when the children were aged 7–8 years, we
located 122 of the stunted children and all the
non-stunted children for a follow up study of
their growth and cognitive development. We
also studied an additional 175 non-stunted
children from the same neighbourhoods who
had been identified during the original survey.17
At the current follow up 116 stunted
children and 190 non-stunted children were
located. Loss to follow up is attributable
primarily to migration. Of the children located,
five did not have their blood pressure measured
because they now lived some distance from
Kingston, giving a sample size of 112 stunted
children and 189 non-stunted children (91.5%
of the children measured at age 7–8 years).
The study was approved by the ethics
committee of the University of the West Indies
and the parents or guardians gave written
informed consent for their children to partici-
pate.
MEASUREMENTS
Anthropometry
The children’s height, weight, head and arm
circumference and triceps and subscapular
skinfolds were measured using standard proce-
dures.18 Interobserver reliability was deter-
mined before beginning the study, and intrac-
lass correlation coeYcients exceeded 0.99 for
all measures. Body mass index (weight/height2,
BMI) was calculated and per cent body fat
estimated from skinfolds with equations for
black children using those for prepubescent
and pubescent children as appropriate.19
Height for age was determined using the
NCHS references.13 Birth weights were ob-
tained from hospital records (73%) or maternal
recall and were available for all of the stunted
children and 181 of the non-stunted children.
Pubertal status
The children’s pubertal status was determined
using the cut oVs recommended by WHO,20
breast stage 2 for girls and genitalia stage 3 for
boys. Children who have reached these stages
have entered their pubertal growth spurt. The
cut oVs were therefore used to classify children
into pre-pubertal and pubertal groups.
Blood pressure
Blood pressure was measured with a mercury
sphygmomanometer after the child had been
sitting relaxed for 10 minutes. Measurements
were taken with the right arm, and arm
circumference was measured to select the
appropriate cuVsize. Eighty nine per cent of
the children were measured using the child/
small adult cuV, and the remaining children
were measured with the adult cuV. Three
measurements were taken, each two minutes
apart, and the average of the last two used. The
first and fifth KorotkoVsounds were recorded.
All measurements were taken by the same
observer. Reliability was measured with an-
other observer in 26 children using a double
headed stethoscope and the intraclass correla-
tion coeYcients were high (systolic r= 0.99;
diastolic r=0.97)
STATISTICAL ANALYSES
Birth weight and current weight were con-
verted to SD scores. Because of the sample
selection procedure, weight at age 9–24
months was bimodal, furthermore, this
measurement was not available for those
non-stunted children enrolled at age 7 years. A
dichotomous variable (stunted/non-stunted)
was therefore used.
Multiple regression analyses were used to
determine the eVects of birth weight, linear
growth retardation and current size on blood
pressure. As recently recommended,21 the
eVects of early size and current size were first
determined separately, then both early and
current size were included in the regression. As
a final step any interactions between early and
current size were investigated to determine
whether early size modified the eVects of
current size. Separate regressions were there-
fore computed for birth weight, current weight
and nutritional status (stunted/non-stunted) at
age 9–24 months. The eVects of adjusting for
current size and any interactions between
earlier size and current size were then deter-
mined. As the stunted children had lower birth
weights than the non-stunted children, birth
weight was included in regressions investigat-
ing the eVects of linear growth retardation at
age 9–24 months to determine its eVects inde-
pendently of size at birth.
Preliminary regressions were conducted for
the stunted and non-stunted groups separately
and for pre-pubertal and pubertal children
separately. The direction and magnitude of the
slopes for birth weight did not diVer signifi-
cantly among the subgroups. The direction of
the slopes for current size was always positive,
however the magnitude of the regression coef-
ficient was greater for the stunted children
compared with the non-stunted children. The
groups were therefore combined to increase
statistical power and an interaction term used
to confirm the diVerence in the eVects of
current weight in stunted and non-stunted
children.
All regression models included current age,
and sex. Current height was included in initial
regressions but was never significant and its
inclusion did not change the findings. Height
was therefore omitted to simplify the models.
Arm circumference, per cent body fat and
pubertal status were oVered stepwise in the
final model.
The intervention in early childhood was a
randomised trial of nutritional supplementa-
tion with or without psychosocial stimulation.14
There were no long term eVects of supplemen-
tation on the children’s size16 and no eVects of
supplementation on blood pressure. The inter-
ventions were therefore not considered further
in the analyses.
Growth and blood pressure 395
www.jech.com
Results
The mean age of the non-stunted children was
slightly greater (approximately 2.5 months, p <
0.001) and they were significantly larger than
the stunted children in all anthropometric
measurements (table 1). In the stunted group
57.1% (n=64) of the children were boys and in
the non-stunted group 53.4% (n=101) were
boys.
The stunted children had caught up in
height relative to the NCHS references, 48.2%
of them now had heights for age>−1SDand
only 11.6% still had heights for age<−2SD.
The BMI distribution of the non-stunted chil-
dren approximated that of the US black popu-
lation22 with 4.2% below the 5th percentile
(thin) and 16.4% above the 85th centile (over-
weight). Few of the stunted children were over-
weight (4.5%) while 18.7% of them were thin
according to the centiles for US blacks.
Significantly more of the non-stunted chil-
dren (59.9%) had entered their pubertal
growth spurt than the stunted children
(39.3%, p < 0.001). More girls (87.5%) than
boys (22.7%, p < 0.001) were in puberty.
The non-stunted children had significantly
higher systolic and diastolic blood pressures
than the stunted children (table 1). These dif-
ferences were no longer apparent after adjust-
ment for age, weight, and height.
PREDICTORS OF BLOOD PRESSURE
Multiple regression analyses of systolic blood
pressure on birth weight and weight at age
11–12 are given in table 2. Birth weight was
significantly related to systolic blood pressure
only after adjustment for current weight. The
eVect of 1 SD change in current weight was 3.8
times that of an SD change in birth weight. The
addition of birth weight to a model containing
age, sex, and current weight resulted in a mod-
est but significant increase in the variance
explained from 30% to 32%. The interaction
between birth weight and current weight was
not significant.
Early childhood growth retardation had no
significant main eVect on systolic blood
pressure after adjustment for birth weight and
current weight (table 3). However, there was a
significant negative interaction between early
childhood growth retardation and current
weight, indicating larger eVects of increasing
weight at age 11–12 years in children who were
stunted at age 9–24 months. There was no sig-
nificant interaction between birth weight and
growth retardation in early childhood.
Arm circumference, per cent body fat and
pubertal status were oVered stepwise after the
variables in the above model were entered.
They made no additional contribution to the
variance in systolic blood pressure.
The same regression analyses were con-
ducted with diastolic blood pressure as the
dependent variable. Birth weight was not
significantly related to diastolic blood pressure
either with or without adjustment for current
weight. Weight at 11–12 years predicted
diastolic pressure and the interaction between
early childhood growth retardation and weight
was also significant (table 4).
Childhood overweight
It has been suggested that the eVect of birth
weight is most strongly expressed in individuals
who become overweight.23 As the number of
Table 1 Age, anthropometry,and blood pressure in stunted
and non-stunted children
Stunted
(n=112)
Non-stunted
(n=189)*
Mean SD Mean SD
Age (y) 11.69 0.34 11.89 0.32
Height (cm) 141.3 6.1 153.4 6.2
Height for age (z score) −1.01 0.83 0.52 0.81
Weight (kg) 32.82 5.77 43.07 8.85
Body mass index 16.35 2.03 18.24 3.22
Arm circumference (cm) 20.2 2.2 22.4 3.0
Body fat (%)† 15.2 6.9 19.0 8.3
Birth weight (kg)‡ 2.90 0.48 3.33 0.52
Systolic bp (mm Hg) 100.4 8.7 106.4 8.5
Diastolic bp (mm Hg)§ 54.1 13.3 59.7 11.3
Adjusted systolic bp (mm Hg)¶ 104.2 9.7 104.1 9.2
Adjusted diastolic bp (mm Hg)§¶ 57.7 15.0 57.5 14.0
*Non-stunted significantly greater than stunted, all p<0.001,
except adjusted systolic and diastolic blood pressures. †Non-
stunted (n=187). ‡Non-stunted (n=181). §n=104 for stunted
group, no measurable diastolic pressure in eight children.
¶Adjusted for age, weight, and height.
Table 2 Multiple regressions of systolic blood pressure at age 11–12 years on birth weight, and weight at age 11–12 years*
â95% CI â95% CI â95% CI â95% CI
Birth weight −0.75 −1.79, 0.29 — −1.28 −2.17, −0.38 −1.27 −2.16, −0.38
Weight at 11–12 years — 4.75 3.82, 5.69 4.90 3.97, 5.83 4.92 3.99, 5.85
Interaction† — — — −0.62 −1.52, 0.29
r20.06 0.30 0.32 0.32
*All models include sex and current age, birth weight and weight at age 11–12 years expressed as SD scores. †Interaction = birth
weight ×weight 11–12 years.
Table 3 Multiple regression of systolic blood pressure on
birth weight, current weight,and nutritional status at age
9–24 months*
â95% CI
Birth weight −1.32 −2.21, −0.42
Weight at 11–12 years 5.22 4.05, 6.39
Stunted/non-stunted at 9–24 months −0.50 −2.93, 1.93
Interaction† −2.61 −5.13, −0.10
r20.33
*Adjusted for sex and current age, birth weight and weight at
age 11–12 years expressed as SD scores. Stunted = 0,
non-stunted = 1. †Interaction = stunted/non-stunted ×weight
at 11–12 years.
Table 4 Multiple regression of diastolic blood pressure on
birth weight, current weight,and nutritional status at age
9–24 months*
â95% CI
Birth weight 0.76 −0.66, 2.17
Weight at 11–12 years 4.18 2.31, 6.05
Stunted/non-stunted at 9–24 months −0.51 −4.33, 3.30
Interaction† −6.28 −10.32, −2.23
r20.13
*Adjusted for sex and current age, birth weight and weight at
age 11–12 years expressed as SD scores. Stunted = 0,
non-stunted = 1. †Interaction = stunted/non-stunted ×weight
at 11–12 years.
396 Walker, Gaskin, Powell, et al
www.jech.com
overweight children was too small for separate
analysis we repeated the analyses excluding the
34 overweight children. Birth weight was still
negatively associated with systolic blood pres-
sure after adjustment for current weight and
the size of the regression coeYcient was little
changed (â= −1.30, 95% CI −2.25, −0.34)
CHANGE IN SYSTOLIC BLOOD PRESSURE
The mean (SD) increase in systolic blood pres-
sure since the previous measurements at age
7–8 years was similar in the stunted (7.9 (7.6)
mm Hg) and non-stunted children (8.0 (8.2)
mm Hg). To look at change in blood pressure
we repeated the regression analyses controlling
for systolic blood pressure at age 7–8 years
(table 5). Children with higher weight at 11–12
years had greater increases in blood pressure
from age 7 to age 11–12 years, while children
with higher birth weight and weight at age 7
years had smaller increases.
Discussion
After adjusting for current size there was no
diVerence in systolic or diastolic blood pressure
between stunted and non-stunted children.
However, the eVect of increased weight at age
11–12 years, on both systolic and diastolic
blood pressure, was greater among children
who experienced linear growth retardation in
early childhood. Thus while there seems to be
no direct relation between stunting and later
blood pressure, stunting may modify the eVect
of later weight gain.
One third of children in developing countries
become growth retarded before the age of 5
years.24 If linear growth retardation potentiates
the eVects of later weight gain on blood
pressure then stunting in early childhood may
be an important contributor to the rise in
cardiovascular disease in developing country
populations undergoing the epidemiological
transition. This may be particularly so where
growth retarded children become overweight
adults.
The expected negative association was ob-
served between systolic blood pressure and
birth weight after adjustment for current
weight.3472526It remains unclear whether this
relation is attributable to size at birth or to the
change in relative size from birth to age 11–12
years.21 The relation between birth weight and
diastolic blood pressure is less consistent, with
negative associations reported,10 25 associations
only in girls27 or boys28 or no associations.52629
No significant relation between birth weight
and diastolic blood pressure was observed in
this study.
The association between birth weight and
systolic blood pressure was seen despite 52% of
the children having entered their pubertal
growth spurt. It has been suggested that rapid
growth during puberty and diVering matura-
tion rates may disrupt the tracking of blood
pressure and for this reason a significant
relation between birth weight and blood
pressure has not been found in some studies of
adolescents29 30 although in one study in which
pubertal stage was assessed, significant inverse
associations were found in pubertal but not
pre-pubertal children.9In the current study
pubertal status did not contribute to the
variance in blood pressure, however, the
children were classified only by whether or not
they had begun their pubertal growth spurt and
this was thus a relatively crude index.
Linear growth retardation in the first two
years of life did not change the eVects of birth
weight on later blood pressure. Whincup et al10
reported a stronger relation between birth
weight and systolic blood pressure at age 3
years in children who were shorter at that time.
Few children in that study would have
experienced the degree of linear growth
retardation (< −2 SD height for age) observed
in children in our study. Further work is
needed on the extent to which postnatal growth
may modify the relation between birth weight
and blood pressure.
The relation between birth weight and blood
pressure was not driven by a stronger relation
in those children who had become overweight
as omitting them for the analyses did not
diminish the regression coeYcient for birth
weight. Thus, unlike the suggestions for
adults,23 our results indicate that the associ-
ation between birth weight and blood pressure
is seen in children within the normal weight
range. The eVect of current weight on systolic
blood pressure was about four times that of
birth weight, similar to findings for children in
the UK aged 8–11 years.27
The increase in systolic pressure between
measurements at age 7 and 11 years was
greater in children with lower birth weight.
Similar findings were reported in children in
the UK measured at age 5–7 years and again at
9–11 years.25 Children who were larger at age 7
years had smaller increases in systolic blood
pressure between ages 7 and 11 years.
Table 5 Multiple regression of systolic blood pressure at
age 11–12 years on birth weight, current weight, and
systolic blood pressure and weight at age 7 years*
â95% CI
Birth weight −1.05 −1.83, −0.27
Weight at 7 years −1.82 −3.29, −0.35
Weight at 11–12 years 4.55 3.08, 6.02
Systolic blood pressure at 7 years 0.49 0.39, 0.58
r20.50
*Adjusted for sex and current age, birth weight and weight at
age 7 years and 11–12 years expressed as SD scores.
KEY POINTS
xChildhood growth retardation may am-
plify the eVect of weight on blood
pressure and contribute to the rise in
cardiovascular disease in developing
countries.
xPostnatal growth retardation did not
change the negative association between
birth weight and blood pressure.
xBlood pressure increased more from age
7–11 years in children with lower birth
weights.
xWeight gain from age 7 to 11 years is
associated with the change in blood pres-
sure during this period.
Growth and blood pressure 397
www.jech.com
Adjusting for weight at age 7 years, a higher
weight at age 11 years was associated with a
larger increase in blood pressure from age 7–11
years. Thus size at birth, in infancy and child-
hood may all be associated with later blood
pressure. Studies of the impact of early growth
on later blood pressure are limited by the
emphasis on birth weight and need to also
include measures of postnatal and childhood
size.
In conclusion, birth weight was a significant
predictor of blood pressure in Jamaican
children, once current weight was controlled.
Postnatal growth retardation may modify the
relation between current weight and blood
pressure. The relation between early growth
and blood pressure is thus not restricted to size
at birth and further studies are needed that
include measures of size during childhood to
fully describe these relations.
Funding: the study was supported by the Wellcome Trust
(Grant no. 049235/Z/96/Z).
Conflicts of interest: none.
1 Barker DJP, Osmond C, Golding J, et al. Growth in utero,
blood pressure in childhood and adult life, and mortality
from cardiovascular disease. BMJ 1989;298:564–7.
2 Barker DJP, Osmond C, Winter PD, et al. Weight in infancy
and death from ischaemic heart disease. Lancet 1989:ii:
577–80.
3 Law CM, Shiell AW. Is blood pressure inversely related to
birth weight? The strength of evidence from a systematic
review of the literature. J Hypertens 1996;14:935–41.
4 Whincup PH. Fetal origins of cardiovascular risk: evidence
from studies in children. Proc Nutr Soc 1998;57:123–7.
5 Donker GA, Labarthe DR, Harrist RB, et al.Lowbirth
weight and blood pressure at age 7–11 years in a biracial
sample. Am J Epidemiol 1997;145:387–97.
6 Godfrey KM, Forrester T,Barker DJP, et al. Maternal nutri-
tional status in pregnancy and blood pressure in childhood.
Br J Obstet Gynaecol 1994;101:398–403.
7 Forrester TE, Wilks RJ, Bennett FI, et al. Fetal growth and
cardiovascular risk factors in the West Indies. BMJ
1996;312:156–60.
8 Law CM, de Sweit M, Osmond C, et al. Initiation of hyper-
tension in utero and its amplification throughout life. BMJ
1993;306:24–7.
9 Zureik M, Bonithon-Kopp C, Lecomte E, et al. Weights at
birth and in early infancy, systolic pressure, and left
ventricular structure in subjects 8 to 24 years. Hypertension
1996;27:339–45.
10 Whincup PH, Bredow M, Payne F, et al. Size at birth and
blood pressure at 3 years of age. The Avon Longitudinal
Study of Pregnancy and Childhood (ALSPAC). Am J Epi-
demiol 1999;149:730–9.
11 Popkin BM. The nutrition transition and its health implica-
tions in lower-income countries. Public Health Nutr 1998;1:
5–21.
12 Barker DJP. Mothers, babies and health in later life. 2nd ed.
Edinburgh: Churchill Livingstone, 1998:97–115.
13 Hamill PVV, Drizd TA, Johnson CL, et al. NCHS growth
curves for children, birth-18 years. Vital and Health Statistics
Series 11, no 165. DHEW publ no (PHS) 78–1650. Hyat-
sville MD: Department of Health and Human Services,
1977.
14 Grantham-McGregor SM, Powell CA, Walker SP, et al.
Nutritional supplementation, psychosocial stimulation and
development of stunted children: The Jamaican study.
Lancet 1991;338:1–5.
15 Walker SP, Powell CA, Grantham-McGregor SM, et al.
Nutritional supplementation, psychosocial stimulation and
growth of stunted children: The Jamaican study. Am J Clin
Nutr 1991;54:642–8.
16 Walker SP, Grantham-McGregor SM, Himes JH,et al. Early
childhood supplementation does not benefit the long-term
growth of stunted children in Jamaica. J Nutr 1996;126:
3017–24.
17 Gaskin P. Early childhood linear growth retardation and its
eVects on blood pressure and lung function in later childhood.
[PhD thesis]. Kingston: University of the West Indies,
Jamaica, 1998.
18 Lohman TG, Roche AF, Martorell R, eds. Anthropometric
standardization reference manual. Champaign IL: Human
Kinetics, 1988.
19 Slaughter MH, Lohman TG, Boileau RA, et al. Skinfold
equations for estimation of body fatness in children and
youth. Human Biol 1988;60:7009–23.
20 World Health Organization. Physical status: the use and
interpretation of anthropometry. WHO Technical report
series 854. Geneva: World Health Organization, 1995:263–
311.
21 Lucas A, Fewtrell MS, Cole TJ. Fetal origins of adult
disease-the hypothesis revisited. BMJ 1999;319:245–9.
22 Must A, Dallal GE, Dietz WH. Reference data for obesity:
85th and 95th percentiles of body mass index (wt/ht2)and
triceps skinfold thickness. Am J Clin Nutr 1991;53:839–46.
23 Leon DA, Koupilova I, Lithell HO, et al. Failure to realise
growth potential in utero and adult obesity in relation to
blood pressure in 50 year old Swedish men. BMJ
1996;312:401–6.
24 United Nations Administrative Committee on Coordina-
tion. Sub-committee on Nutrition(ACC/SCN). Fou rt h
Report on the World Nutrition Situation. Geneva: ACC/SCN/
IFPRI, 2000.
25 Whincup PH, Cook DG, Papacosta O. Birth weight and
blood pressure; cross sectional and longitudinal relation-
ships in childhood. BMJ 1995;311:773–6.
26 Levitt NS, Steyn K, De Wet T, et al. An inverse relation
between blood pressure and birth weight among 5 year old
children from Soweto, South Africa.J Epidemiol Community
Health 1999;53:264–8.
27 Taylor SJC, Whincup PH, Cook DG, et al. Size at birth and
blood pressure: cross sectional study in 8–11 year old chil-
dren. BMJ 1997;314:475–80.
28 Taittonen L, Nuutinen M, Turtinen J, et al. Prenatal and
postnatal factors in predicting later blood pressure among
children: Cardiovascular risk in young Finns. Pediatr Res
1996;40:627–32.
29 MacIntyre S, Watt G, West P, et al. Correlates of blood pres-
sure in15 year olds in the west of Scotland. J Epidemiol
Community Health 1991;45:143–7.
30 Matthes JWA, Lewis PA, Davies DP, et al. Relation between
birth weight at term and systolic blood pressure in adoles-
cence. BMJ 1994;308:1074–7.
398 Walker, Gaskin, Powell, et al
www.jech.com
