SEARCH

SEARCH BY CITATION

Abstract

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. References

Objectives  The clinical characteristics of pre-eclampsia (gestational hypertension and proteinuria) may represent separate pathogenetic conditions. Pre-eclampsia accompanied by restricted fetal growth may originate from abnormal implantation, and appropriate or high birthweights may indicate a mixture of conditions, ranging from mild pre-eclampsia with modest placental involvement to hypertensive conditions without placental disease.

Design  Prospective, observational study.

Setting  General population.

Population  We used data from the Medical Birth Registry of Norway, a population-based registry that has recorded births since 1967. For this study, we used information on length of gestation and presence of pre-eclampsia among 1,679,205 singletons born between 1967 and 1998. Pre-eclampsia was diagnosed in 44,220 (2.6%) pregnancies.

Methods  We studied the risk of pre-eclampsia in relation to standardised measures (z scores) of birthweight, adjusted for length of gestation, and stratified by term and preterm delivery. We also explored whether gestational diabetes was more prevalent in conjunction with preterm than term pre-eclampsia.

Main outcome measures  Pre-eclampsia diagnosed at term or preterm.

Results  For pre-eclampsia diagnosed around term, there was a U-shaped association with birthweight. Compared with appropriate birthweights for gestation, the risk of term pre-eclampsia was more than fourfold higher (relative risk [RR] 4.5, 95% confidence interval [CI], 4.3 to 4.7) if the baby's birthweight was lower than two standard deviations under the mean. For birthweights three standard deviations or higher than the mean, pre-eclampsia was more than twice as likely (RR 2.6, 95% CI 2.2–2.9). In contrast, the risk of preterm pre-eclampsia displayed an L-shaped association with birthweight. Low birthweight (less than −2 standard deviations) was associated with greatly increased risk (RR 9.9, 95% CI 9.1–10.9), but for high birthweights (≥3 standard deviations), there was no association with the risk of preterm pre-eclampsia (RR 1.2, 95% CI 0.7–2.1). The prevalence of gestational diabetes was three times (prevalence ratio 3.3, 95% CI 2.6–3.6) higher in preterm than term pre-eclampsia.

Conclusion  Whereas pre-eclampsia with preterm delivery associated with low birthweight may be caused by underlying placental abnormality, pre-eclampsia delivered at term may represent a mixture of conditions, ranging from mild pre-eclampsia with moderate placental affection to hypertensive conditions in pregnancy without placental dysfunction.


INTRODUCTION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. References

It has been proposed that the common but heterogeneous clinical characteristics of pregnancy hypertension and proteinuria that constitute pre-eclampsia represent separate pathogenetic conditions.1,2 The clinical expression of pre-eclampsia is often classified according to severity, indicated by the extent of proteinuria and blood pressure increase during pregnancy, and by time of delivery, as preterm (before 37 weeks of gestation) or term pre-eclampsia. Whereas most cases of preterm pre-eclampsia originate from an abnormal implantation that causes a shallow connection between the placenta and the endometrium, term pre-eclampsia may represent a mixture of conditions, ranging from moderate placental involvement to a hypertensive reaction to the burden of pregnancy.1–3 Clinically, pre-eclampsia may therefore cover a broad range of underlying categories. For example, the placental abnormality associated with preterm pre-eclampsia is characterised by reduced uteroplacental blood flow that causes placental hypoxia, reduced fetal nutrition, and intrauterine growth restriction.3 Also, metabolic aberrations consistent with cardiovascular risk may be involved in the aetiology of preterm pre-eclampsia.1

However, histopathological studies have shown that placental abnormalities are not always present in conditions that are classified as pre-eclampsia, and that fetal growth may only be reduced if placental function is compromised.4 In term pre-eclampsia, for example, fetal growth often appears unaffected,5,6 and for a small proportion of cases, babies are born large for their gestation.6 Nonetheless, this condition is also characterised by proteinuria and hypertension induced by pregnancy,7 although the underlying pathogeneses may differ depending on whether an abnormality of the placenta is present. Thus, increasing blood pressure and proteinuria towards term may accompany appropriate or even high birthweight for gestational age and be classified as pre-eclampsia.6 Since pre-eclampsia is a clinical diagnosis and placental involvement cannot be reliably determined, misclassification of mild pre-eclampsia at term is a problem that cannot be easily overcome.

In this study, we have distinguished between preterm and term pre-eclampsia, and used data from almost 1.7 million births recorded at the Norwegian Medical Birth Registry. We speculate that preterm pre-eclampsia represents placental disease, and that term pre-eclampsia represents a mixture of conditions, ranging from moderate placental involvement to maternal reactions to the burden of pregnancy. We have explored this possibility by comparing the association of offspring's birthweight with term and preterm pre-eclampsia. We also examined whether gestational diabetes, which is consistent with increased cardiovascular risk,1 is more prevalent in preterm than in term pre-eclampsia.

METHODS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. References

Data were derived from the Medical Birth Registry of Norway that comprises records of more than 1.8 million births between 1967 and 1998.8 Midwives and doctors have to fill in a standardised form to notify the birth registry about each birth that takes place in the country. Central items that are registered include birthweight, birth length, length of gestation, and specific information on complications such as pre-eclampsia, gestational hypertension, and diabetes induced by pregnancy. Birthweight was notified in more than 99% of the births, and gestational age, based on the last menstrual period, was recorded in more than 90% of all births. For this study, we used the information on birthweight and length of gestation to construct standardised z scores for birthweight, adjusted for differences in length of gestation, as previously described.9 Thus, we could construct adjusted z scores for 1,679,205 (92.1%) singleton births among a total of 1,822,982 births.

Pre-eclampsia was defined as an increase in blood pressure to at least 140/90 mmHg after the 20th week of gestation.10 Either the diastolic blood pressure had to be at least 15 mmHg higher than the level measured before the 20th week or the systolic blood pressure had to be at least 30 mmHg higher than the level measured before the 20th week. In addition, proteinuria (protein excretion, at least 0.3 g per 24 hours) had to be present. If present, pre-eclampsia is routinely entered on the standardised form to the birth registry by the midwife or the obstetrician as a specified diagnosis. In some cases, there is information about the presence of hypertension, proteinuria, or oedema during pregnancy. As pre-eclampsia, we included pregnancies with a specified diagnosis of pre-eclampsia and pregnancies with a combination of hypertension induced by the pregnancy and proteinuria.

We identified a total of 44,220 (2.6%) women who had been diagnosed with pre-eclampsia in one of their pregnancies, for whom information on birthweight and length of gestation was available. We categorised pre-eclampsia according to whether the woman was diagnosed before 37 weeks of gestation (preterm) or at term (37–42 weeks of gestation). Gestational diabetes is notified to the Medical Birth Registry on the standardised form filled in at delivery by the obstetrician or the midwife.

We used standardised birthweight, calculated as z scores adjusted for length of gestation as a measure for ‘fetal growth’.9 Briefly, for deliveries at a given week of gestation, sex-specific z scores were estimated as the observed minus the mean birthweight divided by the standard deviation of the mean.9 The z scores were calculated separately within each complete week of gestation in the range of 16–44 weeks for the total distribution of birthweights among live born children for whom reliable information of length of gestation was available.9 We used stratification and logistic regression techniques to estimate odds ratios as estimates of relative risk (RR) and to evaluate possible confounding by other variables. In the analyses, deliveries were divided according to three periods (1967–1976, 1977–1986, and 1987–1998), and maternal age at delivery was divided in five categories (<20 years, 20–24, 25–29, 30–34, and 35 years or more). Age and period were treated as categorical variables in the analyses.

RESULTS

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. References

Among 1,679,205 mothers with singleton births, 44,220 (2.6%) were diagnosed with pre-eclampsia. Of these cases, 37,044 were delivered at term, and the remaining 7176 cases were delivered preterm (before 37 weeks).

The overall relative risk for pre-eclampsia in relation to z scores of birthweight, adjusted for length of gestation and period of diagnosis, displayed a U-shaped association, as shown in Table 1. Compared with the reference category (z score between 0 and 1), the likelihood of pre-eclampsia was nearly six times higher (RR 5.9, 95% CI 5.7–6.2) if the baby's birthweight was lower than two standard deviations under the mean. Simultaneously, it was more than twice as high (RR 2.4, 95% CI 2.1–2.8) if the baby's birthweight was at least three standard deviations higher than the mean.

Table 1.  Relative risk of pre-eclampsia, according to z score of birthweight, adjusted for length of gestation (births in Norway, 1967–1998).
 Birthweight z scores
<−2−2 to −1−1 to 00 to 11 to 22 to 3≥3
  • *

    Adjusted for 10 year period (1967–1976, 1977–1986, and 1987–1998).

Total pre-eclampsia5387960813,43210,21041821149252
Deliveries52,779256,400614,255524,952188,72936,6915399
Risk (%)10.23.82.21.92.23.14.7
Risk ratio5.31.91.11.0 (reference)1.11.62.4
Risk ratio*5.92.011.141.0 (reference)1.131.62.4
95% CI(5.7–6.2)(1.95–2.06)(1.11–1.17) (1.09–1.17)(1.5–1.7)(2.1–2.8)

By stratifying delivery as term (37–42 weeks) or preterm (<37 weeks), we found that the U-shaped association with birthweight was present only for pre-eclampsia with delivery at term (Table 2, Fig. 1). Compared with the reference, the relative risk of term pre-eclampsia was 4.5 (95% CI 4.3–4.7) for birthweights lower than two standard deviations below the mean. For birthweights at least three standard deviations higher than the mean, the relative risk was 2.6 (95% CI 2.2–2.9). For preterm pre-eclampsia, there was a compelling L-shaped association with birthweight z scores (Table 2, Fig. 1). From the reference category, the risk increased successively with decreasing birthweight z score (RR, 1.9, 6.0, and 9.9, respectively). For birthweight z scores higher than the reference, there was no association with preterm pre-eclampsia (RR 1.0, 1.2, and 1.2). Further adjustment for gestational week within the term and preterm categories did not change the results substantially.

Table 2.  Relative risk of pre-eclampsia according to z score of birthweight, adjusted for length of gestation, by preterm and term delivery (births in Norway, 1967–1998).
 Birthweight z scores
<−2−2 to −1−1 to 00 to 11 to 22 to 3≥3
  • *

    Adjusted for 10 year period (1967–1976, 1977–1986, and 1987–1998).

Preterm pre-eclampsia1893240217397403018615
Preterm deliveries772914,27429,40822,72594832359390
Risk (%)24.516.85.93.33.23.73.9
Risk ratio*9.96.01.91.0 (reference)1.01.21.2
95% CI(9.1–10.9)(5.6–6.6)(1.7–2.1) (0.9–1.1)(0.9–1.5)(0.7–2.1)
Term pre-eclampsia3494720611,693947038811063237
Term deliveries45,050242,126584,847502,227179,24634,3325009
Risk (%)7.83.02.01.92.23.14.7
Risk ratio*4.51.611.071.0 (reference)1.141.62.6
95% CI(4.3–4.7)(1.58–1.68)(1.04–1.10) (1.10–1.19)(1.5–1.8)(2.2–2.9)
image

Figure 1. The relation between birthweight z score and risk of pre-eclampsia in preterm and term deliveries.

Download figure to PowerPoint

It has been suggested that the more severe preterm type of pre-eclampsia is associated with factors that are consistent with increased cardiovascular risk.1 We therefore examined whether gestational diabetes was more prevalent in preterm than in term pre-eclampsia, and found that the prevalence of gestational diabetes was roughly three times higher (3.3%versus 1.1%, prevalence ratio 3.1, 95% CI 2.6–3.6) in conjunction with preterm than term pre-eclampsia (Table 3).

Table 3.  Prevalence of gestational diabetes associated with pre-eclampsia (pre-eclampsia in Norway, 1967–1998).
 Diabetes/pre-eclampsiaPrevalence
Prevalence (%)Ratio95% CI
Term398/37,0441.11.0 
Preterm237/71763.33.32.6–3.6

DISCUSSION

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. References

We used data from nearly 1.7 million births to explore the possibility that pre-eclampsia may represent separate pathogenetic entities. We found that low, but not high birthweight was strongly associated with preterm pre-eclampsia, whereas both low and high birthweight was associated with term pre-eclampsia (37–42 weeks of gestation), the latter displaying a strong U-shaped relation.

Previously, a Canadian study has shown that term pre-eclampsia was not associated with reduced birthweight in the offspring, and that a certain proportion of infants with pre-eclampsia born at term were actually large for their gestation.6 These investigators used perinatal information from more than 97,000 births in Canada, which enabled estimates with reasonable precision. A study using data from the Swedish birth registry also found that in mild to moderate pre-eclampsia diagnosed at term, infant birthweight was no different from that of normotensive pregnancies.11 A population-based case–control study of pre-eclampsia in Norway has shown that offspring of clinically severe pre-eclampsia or pre-eclampsia with an early onset (delivery before 34 weeks of gestation) had much lower birthweight than offspring of normotensive pregnancies.5 In that study, offspring of mild and moderate pre-eclampsia with onset after 34 weeks of gestation had birthweights similar to offspring of normotensive pregnancies, and a small proportion of babies in the pre-eclampsia group were large for gestational age. Moreover, analyses of umbilical cord blood showed that insulin-like growth factor-I was low in growth restriction caused by severe pre-eclampsia, but did not differ between offspring of mild to moderate pre-eclampsia and offspring of normotensive pregnancies.12 Since the placenta is an important production site for insulin-like growth factor-I in pregnancy, low cord blood levels may indicate placental dysfunction in these cases.13,14

Our findings support the previous results. The strong association that we found between low birthweight and preterm pre-eclampsia fits well with the hypothesis that pre-eclampsia restricts fetal growth when it is caused by placental abnormalities, resulting in reduced nutrient supply to the fetus.1 This hypothesis also fits with the uteroplacental vascular lesions that have been documented in placental tissues in early onset pre-eclampsia.4,7

We also found that diabetes was diagnosed three times more often in conjunction with preterm than term pre-eclampsia, suggesting a strong association between maternal gestational diabetes and preterm pre-eclampsia. The original hypothesis that pre-eclampsia is a heterogeneous condition suggested that maternal metabolic aberrations consistent with increased cardiovascular risk are involved in the aetiology of more severe, preterm pre-eclampsia.1 Our results appear to be consistent with that hypothesis. Indeed, a possible explanation for our key findings is that women with diabetes (gestational or otherwise) are more likely to have a shallow placental implantation that results in preterm pre-eclampsia and iatrogenic preterm delivery.

The results related to term pre-eclampsia may require more complex explanations. It is possible that some conditions that are classified as pre-eclampsia may represent a mixture of pregnancy conditions that, like pregnancy in general, result in a range of birthweights. High blood pressure and proteinuria in conjunction with labour may be present without placental involvement and signs of systemic disease. Nonetheless, such conditions may be classified as pre-eclampsia and notified as such to the birth registry. Hypertension and proteinuria without placental involvement could also be more prevalent among mothers who carry relatively large babies. If correct, that could explain the U-shaped association between z scores of birthweight and term pre-eclampsia that we found in this study.

Our results indicate that the heterogeneous expression of pre-eclampsia may represent separate pathogenetic entities, instead of being one fundamental process expressing varying degrees of clinical severity. Thus, placental dysfunction is likely to be the basis for pre-eclampsia associated with low birthweight, preterm delivery, and gestational diabetes.1–3,15,16 Pre-eclampsia at term, accompanied by offspring that appears unaffected by the condition, may represent a mixture of conditions, ranging from mild pre-eclampsia with moderate placental involvement to hypertensive conditions without placental dysfunction.

References

  1. Top of page
  2. Abstract
  3. INTRODUCTION
  4. METHODS
  5. RESULTS
  6. DISCUSSION
  7. References