Prospective cohort study of factors influencing the relative weights of the placenta and the newborn infantBMJ 1997; 314 doi: http://dx.doi.org/10.1136/bmj.314.7098.1864 (Published 28 June 1997) Cite this as: BMJ 1997;314:1864
- Lucy A Williams, research fellow in obstetrics and gynaecologya,
- Sharon F Evans, biostatisticiana,
- John P Newnham, clinical professor in maternal fetal medicinea
- a Women and Infants Research Foundation, University of Western Australia, King Edward Memorial Hospital, Subiaco, Western Australia 6008
- Correspondence to: Professor Newnham
- Accepted 15 April 1997
Objectives: To determine the demographic, environmental, and medical factors that influence the relative weights of the newborn infant and the placenta and compare this ratio with other factors known to predispose to adult ill health.
Design: Prospective cohort study.
Setting: The tertiary referral centre for perinatal care in Perth, Western Australia.
Subjects: 2507 pregnant women who delivered a single live infant at term.
Main outcome measures: Placental weight, birth weight, and the ratio of placental weight to birth weight.
Results: By multiple regression analysis the placental weight to birthweight ratio was significantly and positively associated with gestational age, female sex, Asian parentage, increasing maternal body mass index, increased maternal weight at booking, lower socioeconomic status, maternal anaemia, and increasing number of cigarettes smoked daily. There were no consistent relations between the placental weight to birthweight ratio and measures of newborn size.
Conclusions: The ratio of placental weight to birth weight is not an accurate marker of fetal growth. In its role as a predictor of adult disease the ratio may be acting as a surrogate for other factors which are already known to influence health and may act before or after birth. Determining the role that relative growth rates of the fetus and placenta have in predisposing to adult disease requires prospective study to account for the many confounding variables which complicate this hypothesis.
Retrospective analyses have identified an association between a raised placental weight to birthweight ratio and hypertension in adulthood
Accurate estimation of gestational age is crucial when interpreting the placental weight to birthweight ratio
Environmental factors associated with alterations in the placental weight to birthweight ratio may not exert their effects exclusively in the antenatal period
As a marker of fetal growth the potential usefulness of the placental weight to birthweight ratio is diminished because the ratio is influenced by a multiplicity of factors
Prospective study is required to clarify the role of intrauterine programming in the genesis of adult disease
Evidence from large British cohort studies suggests that the combination of a large placenta and low birth weight is a strong independent risk factor for cardiovascular disease in adulthood.1 2 3 This important finding adds to the epidemiological evidence that intrauterine influences are relevant in the development of adult disease.4 5 6 Hypertension is the most important known risk factor for coronary artery disease. Hypertension is significantly more common in people whose history includes a placenta disproportionately large for their birth weight–that is, those who had a high placental weight to birthweight ratio.1
Barker and colleagues formed a broad hypothesis that maternal undernutrition before and during pregnancy influences the relative growth rates of the fetus and the placenta,1 2 3 4 5 6 programming fetal metabolism and predisposing the person to hypertension in adulthood. Animal studies support this hypothesis.7 In human pregnancy both small maternal pelvic diameter (a possible marker of poor nutrition during development) and shortness of babies in relation to head circumference and birth weight (possibly indicating mid-pregnancy nutrient deprivation) are associated with significantly raised placental weight to birthweight ratios.1 Debate has focused on whether these findings represent a true relation or are due to the effects of the many measurable or concealed confounders that are inherent in analyses of this type.8 9
We set out to determine the factors that influence the relative weights of the placenta and the newborn infant and to examine the relation between this ratio and measures of newborn size in order to further evaluate the relevance of this variable as a marker of intrauterine events.
Subjects and methods
The design of the study has been detailed elsewhere.10 Briefly, beginning in 1989 data were collected by questionnaire from 2900 pregnant women and their partners at about 18 weeks of gestation and again at 34 weeks. The study was conducted in Perth, Western Australia, in the sole tertiary level perinatal centre for the state. The potential for introducing bias by using a tertiary referral centre population was minimised by enrolling only women who booked for antenatal care before 18 weeks of gestation, so excluding those referred with complications. Gestational age was ascertained by menstrual dates and ultrasound measurements.
The study was approved by the institutional ethics committee and written consent obtained from each woman at the time of enrolment. Data from the 2507 singleton pregnancies delivered after 37 completed weeks were used in the analysis. Tables 2 3 4 list the variables included. Social class was calculated from a composite score based on the occupation, income, and highest level of education attained by the woman and her partner. Racial groupings11 were categorised by the women and their partners at enrolment. Maternal anaemia was defined as a haemoglobin concentration below 110 g/l at any time in the pregnancy.
Table 1) shows the rate of discordance of gestational age estimation between menstrual dates and ultrasound measurements at 18 weeks. Gestational age was based on the date of the last menstrual period unless there was discordance of more than seven days with ultrasound measurements on or before 18 weeks; in that case the ultrasonic estimate of gestation was used (745 cases; 29.7%). Placental weights were recorded wet without trimming the membranes or cord. Techniques for measurements in newborn infants were as described.12
All data were computerised in an sas format (SAS, Cary, North Carolina). The relations between birth weight, placental weight, and the placental weight to birthweight ratio and various predictor variables were identified by multivariate analysis and logistic regression. To allow for the large number of comparisons probability values of less than 0.01 were considered significant.
Table 2) shows the characteristics of the study population. The mean gestational age at delivery was 39.2 (SD 1.3) weeks. A total of 1205 (48.1%) women were nulliparous and 1434 (57.2%) were in social class I or II. Over a quarter of the population (665 women; 26.5%) smoked during pregnancy, 277 (11.0%) women smoking more than 10 cigarettes a day. Most of the study population was of European origin, the remainder (302 women; 12.0%) being of Asian (predominantly Chinese or Vietnamese), Australian Aboriginal, or Indian origin.
Figure 1) plots placental weight against birth weight. As expected, placental weight and birth weight were highly correlated (r=0.63). The 90% confidence intervals for the individual predicted values showed that in general infants with a higher than average placental weight to birthweight ratio had birth weights and placental weights which spanned the range for typical births.
Table 3 lists the factors which by multiple regression analysis influenced the three outcome variables–that is, birth weight, placental weight, and the placental weight to birthweight ratio. Factors that were significantly and positively associated with the placental weight to birthweight ratio were gestational age at delivery, Asian parentage, female infant, maternal anaemia, a higher maternal weight and body mass index at booking, lower socioeconomic score, and increasing number of cigarettes smoked daily during pregnancy. Maternal prepregnancy weight and maternal height do not act independently of maternal body mass index, and therefore these relations were determined by separate regression analyses.
Table 4) shows the relations between measures of newborn size and placental weight, birth weight, and the placental weight to birthweight ratio. Placental weight and birth weight were significantly greater with larger newborn dimensions and higher ponderal index. However, there were no significant associations between the placental weight to birthweight ratio and length, skinfold thickness, or ponderal index when controlled for gestational age, sex of infant, racial origin, maternal smoking, and maternal body mass index. Ratios of abdominal circumference to head circumference and of length to head circumference were used as indicators of the relative amounts of soft tissue and skeletal growth respectively. Birth weight was strongly and positively correlated with both these measures whereas placental weight was also positively correlated with the abdominal circumference to head circumference ratio but was unrelated to changes in the length to head circumference ratio. The placental weight to birthweight ratio was weakly, positively associated with the ratio of abdominal circumference to head circumference while also being weakly, negatively correlated with the ratio of length to head circumference.
Barker and coworkers together with various other investigators have proposed that poor nutrition before and during pregnancy is a main factor determining the birth weight relative to placental weight and altering fetal metabolism such that there is a predisposition to hypertension in adult life.1 7 13 Sheep fed on poor pastures in mid-pregnancy respond by initially increasing placental growth before any change in fetal growth.7 In human pregnancy a higher placental weight to birthweight ratio has been found in association with women who have lower external conjugate pelvic diameters (indicating possible poor nutrition during development)1 and anaemia (possibly indicating current poor nutritional status).14 15 Placental weight and birth weight are widely available measures with records dating back to at least the early twentieth century. The ratio of these two variables has the potential to be a most useful marker of fetal nutrition and uteroplacental function.
Some of the significant predictors of an increased placental weight to birthweight ratio that we have identified are also environmental risk factors for adult hypertension. Lower socioeconomic status, though not in itself a risk factor for hypertension, is an indicator of lifestyle variables that are risk factors for hypertension–for example, excess alcohol intake and poor dietary patterns. Possibly a high placental weight to birthweight ratio is acting as a surrogate for factors in the maternal environment related to social status to which the growing child may be exposed. We observed that maternal smoking was associated with a significant reduction in birth weight but not placental weight. The effect on birth weight is dose dependent, resulting in an increase in the placental weight to birthweight ratio with the numbers of cigarettes smoked. Other workers who have not found an association between maternal smoking and the placental weight to birthweight ratio did not quantify the numbers of cigarettes smoked.16 Maternal obesity at booking was also positively correlated with the placental weight to birthweight ratio. This suggests that the ratio may be a surrogate for already well recognised risk factors for hypertensive disease related to a family history of obesity and poor dietary patterns. In addition, environmental factors associated with alterations in the placental weight to birthweight ratio may not exert their effects exclusively in the antenatal period.
As expected, placental weight and birth weight were highly correlated. An increased placental weight to birthweight ratio could not be predicted by birth weight or placental weight alone. In particular, a high placental weight to birthweight ratio did not imply an infant small for gestational age. We observed no consistent correlation between the placental weight to birthweight ratio and measures of newborn size. Opposing relations were also found between this ratio and indicators of soft tissue and skeletal growth. The ratio of length to head circumference was weakly and negatively correlated with the placental weight to birthweight ratio, possibly suggesting that a relative reduction in skeletal growth during mid-pregnancy could be associated with an increased risk of hypertensive disease in adulthood. However, the ratio of abdominal circumference to head circumference was weakly and positively correlated with the placental weight to birthweight ratio. If the same reasoning is used this would suggest that poor growth in late pregnancy may be associated with a reduction in the placental weight to birthweight ratio and a possible reduced risk of hypertensive disease in adulthood. In addition to these inconsistencies, there was no significant correlation with ponderal index or length. The pattern which has emerged in these analyses is that associations between the placental weight to birthweight ratio and our measures of newborn size do not present an unequivocal, unidirectional relation. The potential role of this ratio as a marker of fetal growth is thus diminished.
Dangers of extrapolation
We have identified various constitutional and obstetric factors in pregnant women which must be considered when interpreting results of analyses including the placental weight to birthweight ratio. Gestational age has a large influence on the ratio, even within the range of term deliveries. In retrospective studies using old birth records only menstrual dates would have been available, rendering a proportion of the population as unrecognised preterm births. We also observed that girl infants had significantly lower birth weights than boys and only a small trend to lower placental weights, resulting in higher placental weight to birthweight ratios. This highlights that fetal factors are important determinants of fetal growth independent of the placenta. Parental racial origin was also associated with significant differences in weights of newborn infants and their placentas. When both parents were of Asian descent birth weight was significantly lower than that of European offspring, though placental weights were significantly greater. The underlying reason for the increased placental weights in this subgroup is unclear. Analysis of a larger population is required to determine whether this effect is genetic or results from local environmental factors such as diet. This finding may have important implications when extrapolating results from the racially homogeneous population of the United Kingdom at the beginning of this century to the racially heterogeneous population of today.
Extrapolation of findings from populations born at the beginning of this century to those born today is subject to difficulties highlighted by comparison of our cohort with the Preston cohort.1 2 In the 50 years separating the birth dates of the populations birth weights and placental weights significantly increased. Multiple factors probably contributed, including improved maternal nutrition. In addition, the socioeconomic status of our cohort was likely to be higher than that of the Preston cohort because the study required early attendance for antenatal care and because of the comparative affluence of the population of Perth.
The relation between a raised placental weight to birthweight ratio and adult hypertension needs validation for modern births. Potentially modifiable factors have been identified in this analysis. No clear relation, however, has emerged between measures of newborn size and the placental weight to birthweight ratio. From these results we speculate that a raised placental weight to birthweight ratio is acting as a surrogate for other factors already known to influence the development of cardiovascular disease in adult life. Prospective studies in which fetal growth and placental function are measured directly are needed in this exciting topic to determine the role of intrauterine development as an antecedent of adult disease. There is great potential for preventive medicine if a link between modifiable antenatal factors and adult hypertension can be proved. The children in this cohort remain under long term surveillance and their ongoing participation will provide valuable information on the role of intrauterine programming in predisposing to adult disease.
Funding: The Raine Foundation of the University of Western Australia; the National Health and Medical Research Council; and the Women and Infants Research Foundation, Western Australia.
Conflict of interest: None.