- Kirsten Duckitt (), consultant obstetrician1,
- Deborah Harrington, subspecialty trainee in maternal and fetal medicine1
- Correspondence to: K Duckitt
- Accepted 27 January 2005
Objective To determine the risk of pre-eclampsia associated with factors that may be present at antenatal booking.
Design Systematic review of controlled studies published 1966-2002.
Data synthesis Unadjusted relative risks were calculated from published data.
Results Controlled cohort studies showed that the risk of pre-eclampsia is increased in women with a previous history of pre-eclampsia (relative risk 7.19, 95% confidence interval 5.85 to 8.83) and in those with antiphospholipids antibodies (9.72, 4.34 to 21.75), pre-existing diabetes (3.56, 2.54 to 4.99), multiple (twin) pregnancy (2.93, 2.04 to 4.21), nulliparity (2.91, 1.28 to 6.61), family history (2.90, 1.70 to 4.93), raised blood pressure (diastolic ≥ 80 mm Hg) at booking (1.38, 1.01 to 1.87), raised body mass index before pregnancy (2.47, 1.66 to 3.67) or at booking (1.55, 1.28 to 1.88), or maternal age ≥ 40 (1.96, 1.34 to 2.87, for multiparous women). Individual studies show that risk is also increased with an interval of 10 years or more since a previous pregnancy, autoimmune disease, renal disease, and chronic hypertension.
Conclusions These factors and the underlying evidence base can be used to assess risk at booking so that a suitable surveillance routine to detect pre-eclampsia can be planned for the rest of the pregnancy.
Pre-eclampsia is a major cause of maternal and fetal mortality and morbidity.1 2 The incidence of pre-eclampsia is 2-10%, depending on the population studied and definitions of pre-eclampsia.3 With the exception of smoking4 the literature has not been systematically reviewed for factors that predict the relative risk of developing pre-eclampsia. The recent National Institute for Clinical Excellence (NICE) guidelines on antenatal care have reduced the number of antenatal visits recommended for healthy woman at low risk.5 As the randomised controlled trials on which this recommendation was based were never powered to identify important outcomes such as mortality, and as the failure to identify and act on known risk factors at booking contributes to deaths from pre-eclampsia,1 it is important to define risk at the beginning of pregnancy.
We carried out a systematic review of published literature to reach an overall estimate for the risk of pre-eclampsia for each risk factor. This will provide an evidence base from which healthcare professionals can assess each pregnant woman's risk of pre-eclampsia at her booking visit and tailor her antenatal care according to need.
Box 1 Risk factors that can be assessed at booking
Family history of pre-eclampsia
Pre-existing medical conditions:
Insulin dependent diabetes (IDDM)
Time between pregnancies
Body mass index (BMI)
We searched Medline (1966 to July 2002) and Embase (1974 to July 2002) for publications in any language that considered the association between identified risk factors detectable at an antenatal booking visit (box 1) and the subsequent development of pre-eclampsia. A multidisciplinary guideline group set up to formulate evidence based guidelines on the community screening and detection of pre-eclampsia identified the risk factors. Our search terms included pre-eclampsia, preeclampsia, eclampsia, toxaemia, toxemia, meta-analysis, systematic review, risk factors, risk, causality, cohort studies, case-control studies. We also studied reference lists of published letters and classic review articles and asked experts in the field.
We initially selected studies with a cohort or case-control design that included pre-eclampsia (either as an outcome or to define cases) and the risk factors of interest. We assessed the quality of these studies with a checklist, adapted from Taggart et al,6 using participant selection, comparability of groups at baseline, and how the diagnosis of pre-eclampsia was made and according to what definition (box 2). We excluded any study that did not score in any category. We used study size, and the prospective or retrospective design of cohort studies, to score for heterogeneity. We independently extracted data from the included studies and resolved any differences by discussion.
Where data were available, we calculated the unadjusted relative risk with 95% confidence intervals for each study and across studies using the random effects model with the MetaView statistical package (MetaView 4.01, Update Software, Oxford). We calculated the I2> statistic7 for combined studies. This estimates the proportion of total variation in study estimates that is due to heterogeneity and, unlike χ2>, does not depend on the number of included studies. For some risk factors, we could not combine data from the included studies. We have described the results from these studies narratively and presented the published adjusted odds ratios or relative risks separately.
We identified over 1000 studies, and, after screening abstracts, we read 149 papers. We excluded 34 because they were observational studies with no reference group or review articles, nine that reported eclampsia alone or did not separate out preeclampsia from pregnancy induced hypertension, 26 that did not concern the relevant risk factors, and 28 because they scored no points in one or more categories in the quality assessment. Fifty two studies (13 prospective cohort studies, 25 retrospective cohort studies, and 14 case-control studies) were therefore included in the systematic review. Of these, 23 had fewer than 100 participants in at least one of the groups. Table 1 shows details of the quality scores. A list of excluded studies is available on request.
Box 2 Quality assessment of non-randomised studies (points scored)
Selected cohort was representative of the general pregnant population (1)
Cohort was a selected group or the selection of the group was not described (0)
Cases and controls drawn from the same population (1)
Cases and controls drawn from different sources or the selection of groups was not described (0)
Comparability of groups
No differences between the groups explicitly reported (especially in terms of age, parity, pre-existing medical disease, singleton pregnancy) unless it was one of these variables that was under investigation, or such differences were adjusted for (2)
Differences between groups were not recorded (1)
Groups differed (0)
Definition of pre-eclampsia
Referenced definition (2)
Explicit definition that included new onset hypertension after 20 weeks' gestation with new proteinuria (1)
Pre-eclampsia not defined or unacceptable definition (0)
How the diagnosis of pre-eclampsia was made
Review of notes or prospective assignment (2)
ICD or database coding (1)
Process was not described (0)
> 100 participants in each group (2)
< 100 participants in each group (1)
Prospective cohort design (2)
Retrospective design (1)
As expected, there was more evidence of heterogeneity, as assessed by the I2> statistic, in the case-control studies than in the cohort studies. The available published adjusted odds ratios or relative risks, however, were similar to the unadjusted relative risks calculated in the meta-analysis.
Tables 2 and 3 summarise the results by risk factor. We found no data of sufficient quality on the presence of proteinuria at booking. There were also no data to calculate unadjusted relative risks for interval between births, existing hypertension, or existing renal disease. Published adjusted relative risks were available for the interval between births (table 3).
All except one study, which looked at women aged ≥ 40, failed to control or address differences at baseline (particularly pre-existing chronic disease such as hypertension or diabetes). Women aged ≥ 40 had approaching twice the risk of developing pre-eclampsia, whether they were primiparous or multiparous (relative risk 1.68, 95% confidence interval 1.23 to 2.29, and 1.96, 1.34 to 2.87, respectively).8 Nationwide US data suggest that the risk of pre-eclampsia increases by 30% for every additional year of age past 34.9 Young maternal age did not seem to affect the risk of developing pre-eclampsia, whichever cut off age was used.
Nulliparity almost triples the risk for pre-eclampsia (2.91, 1.28 to 6.61) (three cohort studies10–12); this is supported by adjusted odds ratios for nulliparity from two other cohort studies.13 14 Women with pre-eclampsia are twice as likely to be nulliparous as women without pre-eclampsia (2.35, 1.80 to 3.06) (six case-control studies10 15–19).
Women who have pre-eclampsia in a first pregnancy have seven times the risk of pre-eclampsia in a second pregnancy (7.19, 5.85 to 8.83) (five cohort studies12 20–23). Women with pre-eclampsia in their second pregnancy are also more than seven times more likely to have a history of pre-eclampsia in their first pregnancy than women in their second pregnancy who do not develop pre-eclampsia (7.61, 4.3 to 13.47) (seven case-control studies15 16 18 19 24–26).
Family history of pre-eclampsia
A family history of pre-eclampsia nearly triples the risk of pre-eclampsia (2.90, 1.70 to 4.93) (two cohort studies27 28). Women with severe pre-eclamptic toxaemia are more likely to have a mother rather than a mother in law who had had pre-eclampsia.29
When a woman is pregnant with twins her risk of pre-eclampsia nearly triples (five cohort studies, 2.93, 2.04 to 4.21).10 12 18 30 31 Neither the chorionicity nor zygosity of the pregnancies alters this increased risk (data not shown32 33). One study found that a triplet pregnancy nearly triples the risk of pre-eclampsia compared with a twin pregnancy (2.83, 1.25 to 6.40).34
Pre-existing medical conditions
Pre-existing hypertension—In a population based nested case-control study, Davies et al found that the prevalence of chronic hypertension was higher in women who developed pre-eclampsia than women who did not (12.1% v 0.3%).25 McCowan et al compared outcomes in 129 women with chronic hypertension who did not develop superimposed pre-eclampsia with 26 women with chronic hypertension who did.36 Those with superimposed pre-eclampsia had significantly higher rates of perinatal morbidity (odds ratio 8.8, 2.6 to 39.0), small for gestational age infants (5.6, 1.8 to 16.0), and delivery before 32 weeks (15.0, 5.7 to 38.0). A diastolic blood pressure before 20 weeks of either ≥ 110 mm Hg (5.2, 1.5 to 17.2) or ≥ 100 mm Hg (3.2, 1.0 to 7.8) is most predictive of the development of superimposed pre-eclampsia.
Renal disease—Davies et al also found that the prevalence of renal disease was higher in women who developed pre-eclampsia compared with those that did not (5.3% v 1.8%).25 Only one study compared women with renal disease, due to a history of urinary tract infections, with a prospective control population matched for age, parity, smoking, and date of delivery.37 In 69 continuing pregnancies, 6.7% (2/30) of the women who had urinary tract infections developed pre-eclampsia (both primigravida with scarred kidneys) compared with 2.6% (1/39) of women in the control group.
Chronic autoimmune disease—In a matched case-control study Stamilio et al found that women who developed pre-eclampsia were more likely to have an autoimmune disease (relative risk 6.9, 1.1 to 42.3).19
Antiphospholipid syndrome—The presence of antiphospholipid antibodies (anticardiolipin antibodies or lupus anticoagulant or both) significantly increases the risk of developing pre-eclampsia (9.72, 4.34 to 21.75) (two cohort studies38 39). However, when women who developed pre-eclampsia were matched with women who did not, they were no more likely to be positive for lupus anticoagulant or anticardiolipin antibodies (6.12, 0.35 to 108.35) (three case-control studies40 41 42).
Time between pregnancies
In a Norwegian population study Skjaerven et al studied 551 478 women who had two or more singleton deliveries and 209 423 women who had three or more singleton deliveries.43 The association between risk of pre-eclampsia and interval was more significant than the association between risk and change of partner. The risk in a second or third pregnancy was directly related to the time elapsed since the previous delivery. When the interval was 10 years or more the risk of pre-eclampsia was about the same as that in nulliparous women. After adjustment for the presence or absence of a change of partner, maternal age, and year of delivery, the probability of pre-eclampsia was increased by 1.12 for each year increase in the interval (odds ratio 1.12, 1.11 to 1.13).
A cross sectional study from Uruguay found that women with more than 59 months between pregnancies had significantly increased risks of pre-eclampsia (relative risk 1.83, 1.72 to 1.94) compared with women with intervals of 18-23 months.44
A Danish cohort study found that a long interval between pregnancies was associated with a significantly higher risk of pre-eclampsia in a second pregnancy when pre-eclampsia had not been present in the first pregnancy and paternity had not changed.45
Body mass index
Although the studies that looked at body mass index before pregnancy all used different ranges, they all showed effects in the same direction, suggesting an overall doubling of risk of pre-eclampsia with a raised body mass index (2.47, 1.66 to 3.67) (six studies12 16 31 46–48). One cohort study showed that women with a body mass index > 35 before pregnancy had over four times the risk of pre-eclampsia compared with women with a pre-pregnancy body mass index of 19-27 (4.39, 3.52, 5.49).46 We combined all studies that looked at raised compared with normal body mass index at booking and found that the risk of pre-eclampsia is increased by 50%.49–51 Notably, a body mass index > 35 at booking doubles the pre-eclampsia risk (one cohort study, 2.12, 1.56 to 2.88).49 Confounding factors can affect the relation between body mass index and pre-eclampsia as women with raised body mass index may be older and more at risk of chronic hypertension. However, published odds ratios that have been adjusted to take some of these factors into account still suggest an increased risk with a raised body mass index (see table 2). A study comparing low and normal body mass index at booking found that the risk of pre-eclampsia was significantly reduced with a body mass index < 20 (odds ratio 0.76, 0.62 to 0.92, adjusted for diabetes and smoking).52
Blood pressure at booking
Reiss et al matched 30 women with pre-eclampsia for age, race, and parity with normotensive control women.53 Both systolic and diastolic blood pressures were significantly higher in the first tri-mester for women who later developed pre-eclampsia. The study did not define cut off values.
Sibai et al found that higher systolic and diastolic blood pressures at the first visit were associated with an increased incidence of pre-eclampsia (3.8% in women with diastolic blood pressure of < 55 mm Hg, 7.4% in those with diastolic blood pressure 70-84 mm Hg).54 However, their recruitment was limited to women with a first blood pressure reading of ≤ 135/85 mm Hg.
In a population based nested case-control study Odegard et al found that a systolic blood pressure ≥ 130 mm Hg compared with < 110 mm Hg at the first visit before 18 weeks was significantly associated with the development of pre-eclampsia later in pregnancy (adjusted odds ratio 3.6, 2.0 to 6.6).18 The association with a diastolic pressure ≥ 80 mm Hg compared with < 60 mm Hg was similar but not significant (1.8, 0.7 to 4.6).
In a case-control study Stamilio et al found that a mean arterial pressure > 90 mm Hg at the first prenatal visit was significantly associated with the development of severe pre-eclamptic toxaemia (relative risk 3.7, 2.1 to 6.6).19
Confirmed proteinuria at booking
We did not find any studies with an appropriate control group that examined the incidence of pre-eclampsia in women who have proteinuria at booking but no previously known renal disease.
In this systematic review of controlled studies we found that antiphospholipid antibodies, a history of pre-eclampsia, pre-existing diabetes, multiple pregnancy, family history, nulliparity, a raised BMI before pregnancy or at booking, maternal age > 40, renal disease, hypertension, ≥ 10 years since the last pregnancy, and raised blood pressure at booking all increased the risk of a woman developing pre-eclampsia.
We reviewed only published studies, and unpublished studies may contain valid results that conflict with our conclusions. This is of particular concern in a meta-analysis of observational studies as there is a greater tendency towards publication bias than there is with randomised controlled trials.59 Because the peer review process is an important means of ensuring quality, however, possibly only published data and studies should be used.60 Publication bias is always a concern for systematic reviews. Funnel plots for the risk factors where over three studies were included were symmetrical for low maternal age, parity, previous pre-eclampsia, pre-existing diabetes, and body mass index before pregnancy (see figs A-H on bmj.com). Some researchers may not have reported on variables that they studied but that did not show an association with pre-eclampsia.
Pre-eclampsia was seldom divided into early and late onset, nor were results presented for onset of pre-eclampsia or delivery in relation to gestational age. We may therefore have underestimated the importance of risk factors for early onset pre-eclampsia, a type with considerable maternal and perinatal morbidity and mortality.26 61
Although we examined the role of individual risk factors, little is known about the association between them. For instance, is a low risk multiparous woman under 40 who did not have pre-eclampsia in her first pregnancy at an increased risk of pre-eclampsia because she has a family history? Similarly as most of the studies concerning body mass index did not separate out their results for parity or control for previous history, it is also unclear whether a raised pre-pregnancy or booking weight or body mass index is less of a risk factor in a multiparous woman who has not had pre-eclampsia in her first pregnancy.
Because we did not identify any controlled studies of sufficient quality we cannot draw any conclusions about proteinuria at booking, although the association of proteinuria with renal disease, which is a risk factor, suggests that it is probably important.
The risk factors that we have identified can be used to assess risk at the booking visit, so that a suitable surveillance routine to detect pre-eclampsia can be planned for the rest of the pregnancy, as recommended by the recent NICE guideline on antenatal care5 and the new pre-eclampsia community guideline (PRECOG) guideline.62
What is already known on this topic
Various factors, which can be ascertained at the first antenatal or booking visit, are thought to increase the risk of a woman developing pre-eclampsia during pregnancy
What this study adds
The most significant risk factors for developing pre-eclampsia are a history of pre-eclampsia and the presence of antiphospholipid antibodies
Pre-existing diabetes and a pre-pregnancy BMI of ≥ 35 almost quadruple the risk; nulliparity, a family history of pre-eclampsia, and twin pregnancy almost triple the risk; and maternal age ≥ 40, a booking BMI of ≥ 35, and a systolic blood pressure ≥ 130 at booking double the risk
Pre-existing hypertension, renal disease, chronic autoimmune disease, and ≥ 10 years between pregnancies increase the risk but it is not clear by how much
Funnel plots for the main outcomes can be found on bmj.com
We thank Jill Mollison from the Centre for Statistics in Medicine, Oxford who provided statistical advice.
Contributors The original idea for the article came from a meeting of the PRECOG (pre-eclampsia community guideline) development group. The study was designed by KD. Data were extracted independently by KD and DH. The paper was written by KD and reviewed by DH. KD is guarantor.
Funding The photocopying of some of the articles was paid for by the charity APEC (Action on Pre-eclampsia).
Competing interests None declared.
Ethical approval Not needed.