Introduction

During the past decade a series of epidemiological studies in several countries has documented that babies of low birth weight have higher blood pressure as children and adults.^1-3 These relations are maintained after correction for gestational age at birth or other influences on blood pressure in later life, including obesity, socioeconomic group, and alcohol consumption. The association between low birth weight and subsequent hypertension may be explained by inadequate fetal nutrition during pregnancy, resulting either from inadequate maternal nutrition or placental dysfunction.⁴ Fetal adaptations to undernutrition are thought to programme permanent changes in factors that influence blood pressure in later life, including cortisol, growth factors, and insulin.⁴ This model is supported by experiments in animals, in which manipulations in utero, including maternal undernutrition⁵ and treatment with glucocorticoids,⁶ result in lower birth weight, raised blood pressure, and hormonal abnormalities in the adult offspring.

Both blood pressure⁷ and birth weight⁸ are influenced by hereditary factors, so there may be a familial association between low birth weight and hypertension which does not depend on programming of raised blood pressure. Therefore, assessment of the parental background of people with low birth weight and subsequent hypertension is important. Studies which included information about the participants' mothers have shown that maternal smoking, socioeconomic group, and anaemia did not explain the observed relations between low birth weight and subsequent high blood pressure.^9-11 Maternal blood pressure has been assessed during pregnancy, when higher pressures were associated with lower birth weight^12-14 and higher blood pressure in the offspring. ^{13 15-17} Only one study has measured maternal blood pressure outside pregnancy, and this showed that higher blood pressure in 4 year old children was associated with both lower birth weight and higher maternal blood pressure.¹¹ To our knowledge, no study has taken account of any paternal factor other than socioeconomic group. We examined a population of young adults in whom blood pressures have been measured in both parents and sought to establish the contribution of parental blood pressure to the relation between low birth weight and subsequent hypertension.

	Subjects and methods

In 1979, as part of the screening programme for the Medical Research Council's mild hypertension trial, blood pressure was measured in 6088 adults registered at the Ladywell Medical Centre in Edinburgh. In 1986, blood pressure was recorded in 864 young adults registered at the medical centre whose parents had both participated in the earlier screening programme. Some of the 864 offspring were siblings, so the cohort was drawn from 603 families. Details of these offspring and parents have been described elsewhere⁷ and are shown in table 1. All blood pressures were measured in the same way: at any convenient time of day, after 5 minutes of sitting, and by two trained observers using a Hawksley random zero sphygmomanometer. The means of the two readings were used for analysis. Offspring were also weighed and their height was measured, and women were asked if they were using an oral contraceptive. Offspring from this cohort were then selected from groups with maximum contrast in their predisposition to high blood pressurethat is, on the basis of offspring and parental blood pressures being in the upper or lower third of the age adjusted distribution. Subgroups of these offspring participated in a series of further studies, which are reviewed elsewhere. ^{18 19}

Collection of data on birth weight
In 1994 all of the 603 mothers were sent letters asking them to recall the birth weight of their offspring. Replies were received from 398 mothers, providing data for 545 offspring. In addition, the case records at Ladywell Medical Centre were examined to determine the birth weights recorded by midwives. This provided birth weights for 349 offspring, 259 of whom had birth weights recalled by their mother. We validated the maternal recall of birth weights by comparing their values with the birth weights recorded by the midwives. As described in other populations, ^{20 21} maternal recall of birth weight was accurate (95% confidence interval of the mean difference between birth weights recalled by mothers and recorded by midwives was 32 g to 22 g; the 95% confidence interval for an individual measurement, or prediction interval, was 467 g to 462 g).²²

Statistical analysis
Stepwise linear regression analysis was used to identify correlates of birth weight and systolic and diastolic blood pressure in the offspring. All observed relations were stronger for systolic than diastolic blood pressure, so only systolic blood pressures were used in further analyses. Data are shown as means (SD).

	Results

The blood pressure of the offspring was higher in men than women and positively correlated with current body weight (table 2). In addition, higher maternal or paternal systolic blood pressure was associated with higher systolic blood pressure in the offspring. Birth weight was lower in the offspring of younger mothers (table 3). In addition, higher maternal, but not paternal, systolic blood pressure was associated with lower birth weight in the offspring.

If parental blood pressures were ignored birth weight and subsequent systolic blood pressure were inversely correlated after correction for sex and current body weight (P=0.06), such that a 1 kg decrease in birth weight was associated with a 2.24 mm Hg increase in systolic blood pressure (table 2). However, if parental blood pressures were included in the regression model they had a significant effect, such that the relation between birth weight and blood pressure was weaker (P=0.15), a 1 kg decrease in birth weight being associated with a 1.71 mm Hg increase in systolic blood pressure (table 2).

	Discussion

This study confirms a relation between low birth weight and high blood pressure in young adults which is of similar strength to that found in other groups of this age.¹ Given the small sample size relative to other studies, this relation was not significant. The study also confirms the well documented influence of parental blood pressure on blood pressure in offspring.⁷ Our study allows the contribution of parental blood pressure to the relation between low birth weight and high blood pressure in adult offspring to be assessed. It shows that mothers with higher blood pressure measured between 9 and 19 years after their pregnancy had offspring with lower birth weight (P<0.01). These results suggest that low birth weight is, at least in part, a feature of the inherited contribution to hypertension. In this context, the inherited contribution may be mediated by genetic or shared environmental factors, or both. As a result, the correlation between low birth weight and subsequent high blood pressure is confounded by the influence of parental blood pressure.

The size of this confounding effect is difficult to measure because of differences in the precision of the measurements. Both inherited and non-inherited factors are likely to contribute to the relation between birth weight and systolic blood pressure. When we accounted for parental blood pressure, the slope of the relation was reduced from 2.24 mm Hg/kg to 1.71 mm Hg/kg (table 2), suggesting that as much as 24% is accounted for by inherited factors. However, the offspring in this study, born in Edinburgh during 1959-70, lived in a comparatively favourable economic and nutritional environment. The influence of variations in maternal nutrition during pregnancy may be comparatively small in this group, making the inherited influence more apparent. In other populationsfor example, those born in India or in wartime Europe, ^{23 24} nutritional factors may be more important. None the less, these observations have important implications for attempts to explain the correlates of low birth weight in humans. In addition to investigation of programmed responses to events in early life, inherited mechanisms should be sought which could explain the association between low birth weight and high blood pressure.

Contrasting influence of mother's and father's blood pressure
Birth weight was related to maternal but not paternal blood pressure. This may be because of the influence of maternal size on birth weight⁸ and the putative imprinting of exclusively maternal genes.²⁵ However, a more likely explanation is that birth weight is dependent on blood pressure during pregnancy. Blood pressure during pregnancy is related to that in later life.²⁶ Mothers with higher blood pressure seemed to deliver smaller babies only if the hypertension was severe and resulted in premature delivery. ^{14 16} Therefore, maternal blood pressure was thought not to contribute to the association between low birth weight and adult hypertension which occurs across the range of birth weight and is independent of gestational age. ^{1 4} We were not able to obtain reliable data for gestational age in our subjects to establish whether, across the range of normal blood pressure, mothers with higher blood pressure delivered earlier. However, recent data using ambulatory blood pressure monitoring has shown that higher maternal blood pressure is associated with lower birth weight independently of gestational age.¹² The current observations could be explained if mothers who have higher blood pressure in middle age also had higher blood pressure in pregnancy which was associated with lower birth weight in their offspring.

Conclusion
This study shows a link between two major areas of research in hypertension: the fetal origins and hereditary hypotheses. It will be important to establish whether other correlates of low birth weightfor example, insulin resistanceare subject to the same confounding interactions.