Birth weight and blood pressure: cross sectional and longitudinal relations in childhoodBMJ 1995; 311 doi: https://doi.org/10.1136/bmj.311.7008.773 (Published 23 September 1995) Cite this as: BMJ 1995;311:773
- Peter Whincup, senior lecturer in clinical epidemiologya,
- Derek Cook, senior lecturer in epidemiologyb,
- Olia Papacosta, research statisticiana,
- M Walker, research administratora
- a University Department of Public Health, Royal Free Hospital School of Medicine, London NW3 2PF
- b Department of Public Health Sciences, St George's Hospital Medical School, London SW17 0RE
- Correspondence to: Dr Whincup.
- Accepted 14 July 1995
Objective: To examine cross sectional and longitudinal relations between birth weight and blood pressure in childhood.
Design: Cross sectional study of primary school children aged 9-11 years, with analysis in relation to previous measurements at 5-7 years in a subgroup.
Setting: 20 primary schools in Guildford and Carlisle.
Subjects: 1511 children measured at 9-11 years (response rate 79%), including 549 who had been measured at 5-7 years.
Main outcome measures: Blood pressure at 9-11 years, change in blood pressure between 5-7 and 9-11 years, birth weight (based on maternal recall), and placental weight (based on birth records).
Results: At 9-11 years birth weight was inversely related both to systolic blood pressure (regression coefficient −2.80 mm Hg/kg; 95% confidence interval −3.84 to −1.76) and to diastolic blood pressure (regression coefficient −1.42 mm Hg/kg; −2.14 to −0.70) once current height and body mass index were taken into account. Placental weight was inversely related to blood pressure after adjustment for current height and body mass index but placental ratio (placental weight to birth weight) was unrelated to blood pressure. Between 5–7 and 9–11 years systolic blood pressure rose more rapidly in children of lower birth weight (regression coefficient −1.71 mm Hg/kg; −3.35 to −0.07). This effect seemed to be stronger in girls.
Conclusions: Birth weight rather than placental ratio is the early life factor most importantly related to blood pressure in childhood. The results support the possibility of “amplification” of the relation between birth weight and blood pressure, particularly in girls.
The placental ratio (placental weight to birth weight) is not related to blood pressure in children
The relation between birth weight and blood pressure becomes stronger (that is, “amplified”) with increasing age in childhood
Whether amplification provides evidence of “programming” has still to be resolved
Barker and colleagues argued that early life factors, acting particularly in utero, may have a strong influence on subsequent cardiovascular risk.1 An inverse relation between birth weight and subsequent blood pressure (lower birth weight being associated with higher subsequent mean blood pressure) has been reported in adults2 3 4 and children,2 5 6 7 8 though not consistently in adolescents.9 10 11 Earlier reports also suggested that a large placenta, particularly in the presence of low birth weight, was related to subsequent high blood pressure and the risk of hypertension.3 Recent reports have suggested that the birth weight and blood pressure relation becomes stronger with age (“amplification”)12 and that this may be an example of “programming.”13 However, the evidence for amplification is based almost entirely on comparison of cross sectional studies with subjects of different ages and times of birth rather than on longitudinal measurements.14
We studied the relation between birth weight and blood pressure in children. In cross sectional analyses at 9-11 years of age we examined the relations between birth weight and blood pressure and the influence of placental weight. In longitudinal analyses based on children measured at both 5-7 and 9-11 years of age we examined the development of the birth weight and blood pressure relation with age.
Subjects and methods
We present data based on two surveys four years apart (1987 and 1991) in Guildford and Carlisle. The sampling and methods used have been detailed elsewhere.6 15 16 Both surveys were carried out in the same sample of 10 primary schools in each town. The second survey included all children from the earlier survey still attending the schools together with similar numbers of children in the same classes. In both surveys measurements were made by a team of trained nurse observers, who visited schools in the two areas in sequence.
Blood pressure was measured after five minutes' rest with the child seated and the arm supported at chest level. Three blood pressure measurements were made one minute apart on the right arm with the Dinamap 1846SX oscillometric blood pressure recorder (Critikon Inc, United States).17 All analyses were based on the means of the three readings. At 9-11 years all measurements were made with the adult size cuff (cuff bladder 22 cm × 12 cm); at 5-7 years all measurements were standardised to the child size cuff (cuff bladder 15 cm x 9 cm), as described.18 These cuff sizes ensured that the minimum cuff bladder width to arm circumference ratio of 40% recommended by the American Heart Association19 was met for 98% of the study population at age 9-11 years and for 99% of the population at age 5-7 years. Ethnic group was assessed on the basis of the child's appearance.
A self administered parental questionnaire sent to the parents on the day of examination in the 1991 survey was used to collect information on birth weight and gestational age (assessed on the basis of recall of the number of weeks the birth was early or late). With the permission of mothers, hospital birth record information was collected on subjects born in the two study areas, including birth weight, gestational age based on date of last menstrual period, and (available in Carlisle only) placental weight. Because there was close agreement in birth weight and gestational age between maternal recall and birth record information, analyses of birth weight and blood pressure at 9-11 years were based on maternal recall data.
A total of 924 children aged 5-7 years (84% response rate) took part in the 1987 survey; 1511 children aged 9-11 years (79% response rate) took part in the 1991 survey. All but eight children were of European origin; exclusion of the others had no effect on the results. Among the 1511 children measured in 1991, questionnaires with full birth details were returned for 1230 (81%). Of these, 996 children were born in the study towns; maternal consent to retrieval of birth records was obtained for 995 children. Complete birth records were obtained for 838 (84%) of these children, 473 of whom were Carlisle subjects with data on placental weight. The 1991 survey included 549 children of 699 (79%) who took part in the 1987 survey and were still attending the survey schools. The longitudinal relations between birth weight and blood pressure were examined in 523 of these 549 subjects for whom maternal recall information on birth weight was available.
Statistical analysis--All analyses were carried out with the statistical analysis software (SAS) package.20 Birth weight, placental weight, age, and measures of current body size--height and body mass index (weight (kg) divided by height --were either treated as continuous variables or ranked and grouped in equal thirds or fifths. Standard linear regression techniques were used throughout and all analyses standardised for the effect of study area. Changes in blood pressure between 5-7 and 9-11 years were examined both as absolute changes and as changes in rank by using standard deviation scores. These were calculated separately for blood pressure at 5-7 years and 9-11 years, standardised for age, sex, height, and body mass index at each age. Changes in standard deviation scores were then regressed on birth weight to examine the relation between birth weight and change in blood pressure rank between 5-7 and 9-11 years.
BIRTH WEIGHT AND CURRENT BODY SIZE (AGE 9-11)
Birth weight showed little or no relation with blood pressure in a model adjusted only for age and sex (fig (1). However, after additional adjustment for height and body mass index (correlation coefficients with birth weight r=0.19, r=0.10 respectively) birth weight showed a smooth graded inverse relation with both systolic and diastolic pressure (fig (1). The corresponding regression coefficients defining the increase in blood pressure for a 1 kg fall in birth weight were −2.80 mm Hg (95% confidence interval −3.84 to −1.76; P<0.0001) for systolic pressure and −1.42 mm Hg (−2.14 to −0.70; P<0.0001) for diastolic pressure. The relations between birth weight and blood pressure were very similar in preterm and term infants, in boys and girls, and in children in the top and bottom thirds of the distribution of height and body mass index. They were also independent of parity, social class, maternal blood pressure, and maternal smoking in pregnancy.
BIRTH WEIGHT AND PLACENTAL WEIGHT (AGE 9-11)
Relations between birth weight, placental weight, and blood pressure were examined in 473 Carlisle children with data on these variables. Placental weight was strongly correlated with birth weight (r=0.45). Like birth weight, it showed a smooth inverse relation with blood pressure, particularly systolic pressure, once current height and body mass index were taken into account. For a 100 g fall in placental weight systolic pressure rose by −0.84 mm Hg (95% confidence interval −1.50 to −0.18; P=0.012). However, including birth weight in the regression model greatly attenuated the relation between placental weight and blood pressure. Moreover, the placental ratio (placental weight to birth weight) showed no relation either with systolic pressure (P=0.89) or with diastolic pressure (P=0.37).
When the mean systolic blood pressure for each third of placental weight and birth weight was examined (table I) the highest mean blood pressure was observed in the group in the lowest third of both birth weight and placental weight distributions; subjects with low birth weight and high placental weight did not have a particularly high mean blood pressure. Similar findings were recorded for diastolic pressure (data not shown). There was no evidence of interaction between birth weight and placental weight, either for systolic (P=0.56) or diastolic (P=0.70) blood pressure. These results were unaffected by exclusion of preterm births (</= 36 weeks).
CHANGES BETWEEN 5-7 AND 9-11 YEARS
The relative strengths of the birth weight and blood pressure relation at 5-7 and 9-11 years, standardised for age, sex, height, and body mass index at each age, were examined in 523 children studied longitudinally and in whom birth weight data were available (table II). The regression coefficients, defining the difference in blood pressure for a 1 kg increase in birth weight, increased between 5-7 years and 9-11 years by 73% for systolic pressure and 22% for diastolic pressure. The corresponding within subject analysis taking into account age, sex, height, and body mass index at both ages showed that the increase in the strength of the systolic relation was significant whereas that for diastolic pressure was not. The increase in the strength of the systolic relation appeared to be more pronounced in girls (regression coefficient −3.42 (SE 1.13) mm Hg/kg) than in boys (regression coefficient 0.03 (1.13) mm Hg/kg) (test for interaction, P=0.03).
The extent to which birth weight was related to change in blood pressure rank between 5-7 and 9-11 years was examined by determining the standard normal deviate of blood pressure in each subject at 5-7 and 9-11 years separately (taking into account age, sex, height, and body mass index at each age) and studying the relation between birth weight and change in standard normal deviate (table II). Overall, birth weight had little or no relation to change in the rank of blood pressure between these ages. However, birth weight seemed to be inversely related to a change in blood pressure rank in girls (regression coefficient −0.26 (SE 0.12)) rather than in boys (regression coefficient 0.09 (0.12)) (test for interaction, P=0.05).
The inverse relation between birth weight and blood pressure after standardisation for current body size was consistent with the findings of earlier studies in children.2 5 6 7 8 The associations seen at 9-11 years were similar to those observed by Cater and Gill5 but considerably stronger than those reported by other workers.2 12 The association was smoothly graded, with no evidence of the U shaped relation suggested by Launer et al.21 The results are consistent with our earlier report in suggesting that attained birth weight is the important factor in the relation between birth weight and blood pressure,8 regardless of the degree of intrauterine growth retardation.
PLACENTAL WEIGHT AND BLOOD PRESSURE
Our results do not support the hypothesis that placental weight or the ratio of placental weight to birth weight is a useful marker of higher subsequent blood pressure. In this study, as in an earlier, small study of 7 and 18 year olds,22 placental weight was inversely related to blood pressure. However, placental weight showed little relation with blood pressure once birth weight was taken into account, and placental ratio was unrelated to blood pressure.
CHANGE IN RELATION WITH AGE
This is one of the first longitudinal studies to examine the way in which the birth weight and blood pressure relation develops with age directly--an issue of direct relevance to the possibility of programming.13 Our results suggest that the inverse relation between birth weight and blood pressure emerges during the first five years after birth and that the relation becomes stronger (that is, “amplified”) between 5-7 and 9-11 years. The increasing strength of the relation is not an artefact of cuff size or of other aspects of the measuring procedures used. However, it is observed only for systolic pressure and seems to be more pronounced in girls. This is surprising given the greater physical maturity of girls at this age and the inconsistent findings of earlier studies of the birth weight and blood pressure association in adolescence.9 10 11 However, the greater strength of the relations between birth weight and blood pressure at 9-11 years in this study compared with earlier studies2 12 emphasises the importance of substantiating the finding in other longitudinal studies.
It is also important that the relation between amplification and “programming” is defined. In particular, agreement is needed on whether amplification alone is evidence of programming or whether a particular mechanism of amplification is required. Theoretically, the birth weight and blood pressure relation could be amplified simply by blood pressure tracking combined with the dispersion of the blood pressure distribution which occurs with age (fig2). Whether amplification occurring by this mechanism (option (a)) is evidence of programming or whether evidence of systematic upward “detracking” across the blood pressure distribution by children of low birth weight (option (b)) is required (fig 2) has still to be established. Our present results provide some support for the possibility of detracking--though only in girls--and emphasise the need for further longitudinal studies of the relation between early life factors and cardiovascular risk factor development.
Two of us (PW and MW) were supported by the British Heart Foundation and one (OP) by the Stroke Association. We thank the research team (Jane King, Susan Wright, Joan Branthwaite) and the education authorities, schools, parents, and children for their cooperation and local directors of public health and hospital medical records departments for their help.
Funding Medical Research Council and Coronary Artery Disease Research Association.
Conflict of interest None.