Risk factors and clinical manifestations of pre-eclampsia
Correspondence: Dr R. Ødegård, Institute of Cancer Research and Molecular Biology, University Medical Center, N-7489 Trondheim, Norway.
Objective To study associations between established risk factors for pre-eclampsia and different clinical manifestations of the disease.
Design A population-based, nested case-control study.
Setting Information from 12,804 consecutive deliveries that took place over three years at a birth clinic, which alone serves a population of nearly 240,000 in Rogaland county, Norway.
Subjects Cases of pre-eclampsia (n= 323) and healthy controls (n= 650) were selected. Pre-eclampsia was defined as increase in diastolic blood pressure (≥ 25 mmHg to ≥ 90 mmHg) and proteinuria (≥ 1+ by dipstick testing) after 20 weeks of pregnancy.
Main study measures Parity, previous pre-eclampsia, blood pressure, maternal weight, and maternal smoking were included as study varstudy variables. Women with pre-eclampsia were grouped according to clinical manifestations of the disease (i.e. severity [mild, moderate or severe]) and time of onset (early or late gestation). Associations with the study factors were estimated as relative risks (odds ratio, OR).
Results Both nulliparity and hypertension increased pre-eclampsia risk, with no clear preference for any clinical subtype. High maternal weight was related to a higher risk of mild and moderate, but not severe, pre-eclampsia. Previous pre-eclampsia strongly increased the risk for pre-eclampsia in the current pregnancy, and the risk of early onset disease was especially high (OR 42.4; 95% CI 11.9–151.6). Overall, smoking was associated with a reduced risk for pre-eclampsia (OR 0.6; 95% CI 0.4–0.9). However, no effect of smoking was observed in the early onset disease group and among women with repeated pre-eclampsia.
Conclusion Nulliparity and hypertension increased the risk for each subgroup of pre-eclampsia, but high maternal weight, previous pre-eclampsia and smoking were not consistently associated with each clinical subtype. This observation may suggest that heterogeneous clinical manifestations of pre-eclampsia may be preceded by different pathological mechanisms.
Pre-eclampsia is a major contributor to maternal and fetal mortality and morbidity1. The incidence of preeclampsia ranges from two to ten percent between studies, depending on the population studied and the criteria used to diagnose pre-eclampsia2–4. The basis for the diagnosis is an increase in blood pressure during pregnancy, but the extent of increase that is required for diagnosis varies between studies. For the most often used diagnostic criteria proteinuria should also be present, but in some studies this has not been a requirement. The fact that several classification systems for pre-eclampsia are in current international use5 complicates comparisons between different studies.
Shallow endovascular cytotrophoblast invasion in the spiral arteries and generalised maternal endothelial dysfunction are characteristics of pre-eclampsia, but its aetiology remains obscure6. Genetic, immunologic, environmental and vascular-mediated factors are all thought to play an important role in the development of preeclampsia7,8, and several risk factors have been identified. Among those, nulliparity, previous pre-eclampsia, high maternal weight, hypertension, diabetes and twin pregnancy9–11 have shown a consistent association with increased risk. In most analyses cases have encompassed all clinical manifestations of pre-eclampsia, failing to distinguish between cases of dramatic early onset and cases with mild symptoms gradually developing towards term.
A reduced risk of pre-eclampsia in smoking women has been observed in several studies3,12, whereas others have found no such effect4. Tobacco smoke contains vasodilating components13, and smoking may have a hypovolaemic effect14. It has been shown that smokers may have lower blood pressure than non-smokers15 and that the incidence of gestational hypertension is lower among pregnant smokers16. Thus, observations may suggest that smoking prevents pre-eclampsia by lowering blood pressure. There is little research on the effect of smoking related to different clinical subgroups of pre-eclampsia, although one study has reported that the risk for both mild and severe preeclampsia is similarly reduced12.
The heterogeneous manifestations of pre-eclampsia may be preceded by a variety of mechanisms. Thus, insight into how established risk factors for pre-eclampsia interfere with the development of different clinical subtypes, may provide a clue to a better understanding of the heterogeneous pathogenesis17. In this study we have examined the association with risk factors for preeclampsia in general and explored characteristics of risk factor patterns between different clinical subtypes of pre-eclampsia. We have used a relatively strict definition of pre-eclampsia18 to estimate its incidence in a population of nearly 13,000 pregnant women. Antenatal factors were prospectively recorded before 18 weeks of gestation, and analysed in relation to subsequent risk of pre-eclampsia among nulliparous and parous women, according to disease severity and time of disease onset.
This case-control study from a defined region in Norway covers all births from January 1993 to December 1995 at the Central Hospital in Rogaland County, Stavanger, a region of 238,806 inhabitants. The Central Hospital is the only delivery unit in this region, and in all 12,804 deliveries took place during the study period. The study has been approved by the Regional Committee for Ethics in Medical Research, and study subjects were included after written consent was obtained.
Cases and controls
The Norwegian Medical Birth Registry has used standardised forms to record information on all deliveries since 196719. To identify study subjects, we followed three procedures. First, the Medical Birth Registry was used to identify all women registered with pre-eclampsia who delivered at Rogaland Central Hospital. Further potential cases were identified from consecutive records of midwives at the delivery station, and thirdly, we identified all pregnancies with “proteinuria” or “hypertension” from computerised hospital data. In all, 1300 women (10%) were identified as potential cases. We reviewed the hospital records for each potential case, and included women as cases of pre-eclampsia if they fulfilled the following criteria.
Hypertension and proteinuria were defined according to the CLASP trial18. Briefly, persistent diastolic blood pressure of at least 90 mmHg had to develop after 20 weeks of gestation, in addition to an increase in diastolic blood pressure of at least 25 mmHg. Women with a baseline diastolic blood pressure of 90 mmHg or higher were included as cases if diastolic blood pressure had increased by at least 15 mmHg. Proteinuria was defined as 0.3 mg/L (semiquantitative dipstick 1+) in at least one urine sample after 20 weeks of gestation without simultaneous urinary infection. Twelve women with no history of hypertension had missing blood pressure before 20 weeks of gestation, but had a diastolic blood pressure of ≥ 105 mmHg at a later visit, and it was decided to include these women as cases of pre-eclampsia.
Using these criteria, we identified 323 cases of preeclampsia. We categorised pre-eclampsia into mild, moderate or severe disease, as suggested by Redman20. Mild pre-eclampsia was defined as an increase in diastolic blood pressure of ≥ 25 mmHg and proteinuria 1+ on semiquantitative dipstick, moderate pre-eclampsia as an increase in diastolic blood pressure of ≥ 25 mmHg and proteinuria 2+ on semiquantitative dipstick, and severe pre-eclampsia as diastolic blood pressure of ≥ 110 mmHg, increase in diastolic blood pressure of ≥ 25 mmHg, and proteinuria 3+ on semiquantitative dipstick, or at least 500 mg/24 hours.
Haemolysis, elevated liver enzymes and low platelet syndrome or eclampsia was interchangeable with severe pre-eclampsia, and eclampsia included seizures in addition to hypertension and proteinuria. Haemolysis, elevated liver enzymes and low platelet syndrome was defined as elevated serum liver enzyme concentrations (aspartate aminotransferase and/or alanine aminotransferase > 70) or low platelet counts (≤ 100,000/ml), in addition to having epigastric pain and pre-eclampsia. Pregnancy termination before or at 32 weeks of gestation was treated as a proxy variable for early onset pre-eclampsia.
The Medical Birth Registry selected women as controls from the population of pregnant women who delivered at Rogaland Central Hospital. Approximately half the controls were women who delivered subsequent to a case, and the other half were frequency matched to the case group on maternal age. Thus, the case-control study consisted of 323 cases of pre-eclampsia and 650 women without pre-eclampsia.
Antenatal information was collected before 18 weeks of gestation on standardised forms. Smoking was divided into three categories: non-smokers; smoking 1–4 cigarettes per day; and smoking ≥ 5 cigarettes per day. Maternal weight and blood pressure were each divided into four categories: maternal weight in four steps of 10 kg per unit with < 60 kg as reference category, and diastolic and systolic blood pressure were categorised using an increment of 10 mmHg per unit, starting at 60 mmHg and 110 mmHg, respectively. Prepregnancy diseases and parity status were also recorded. Prepregnancy diabetes, a well-known risk factor for pre-eclampsia, was not analysed because it was present only for five cases and one control.
Our main aim was to study the association between the study variables mentioned above and the risk of preeclampsia. We estimated odds ratios as a measure of relative risk, and used unconditional logistic regression to adjust for potentially confounding effects in the multivariable analysis21. Where appropriate, we estimated linear effects as trends across ordinal categories of the given variable. Results are given as OR, with the corresponding 95% CI. All statistical analyses were done using SPSS version 7.5.
In this population of 12,804 pregnant women, the risk of pre-eclampsia was 2.5%. Pre-eclampsia was diagnosed among 208 nulliparous (64%) and 115 parous women (36%), and 51 of the parous women (44%) had a history of pre-eclampsia in a previous pregnancy. Among 323 cases, 107 (33%) were mild, 130 (40%) were moderate, and 86 (27%) had severe disease. According to onset of symptoms, 10.5% were classified as early onset preeclampsia, and these cases were dominated by severe disease, compared to cases with late onset (Table 1). Eclampsia occurred among 6 (2%), yielding a risk of 0.05%. Haemolysis, elevated liver enzymes and low platelet syndrome was diagnosed among 16 (5%), which corresponds to 0.1% of the total population.
Table 1. Early and late onset of pre-eclampsia (PE), according to clinical severity. Values are given as %.
Women with pre-eclampsia in a previous pregnancy had a strongly increased risk of pre-eclampsia (Table 2) in the current pregnancy compared to parous women with no previous pre-eclampsia (OR 21.5, 95% CI 9.8–47.2). The association was particularly strong for moderate and severe disease (Table 3), and extremely strong (OR 42.4, 95% CI 11.9–151.6) for early onset disease (Table 4).
Table 2. The association (odds ratio) between risk factors recorded at first antenatal visit (before 18 weeks of gestation) and subsequent risk of pre-eclampsia (PE). Values are given as n (%) or OR [95% CI], unless otherwise indicated.
|PE in a previous pregnancy*|| || || || || |
| No||64 (56)||394 (97)||1.0||1.0|| |
| Yes||51 (44)||13 (3)||24.2||21.5 [9.8–7.2]|| |
|Nulliparous|| || || || || |
| No||115 (36)||407 (64)||1.0||1.0|| |
| Yes||208 (64)||232 (36)||3.2||3.6 [2.6–5.0]|| |
|Blood pressure (mmHg)|| || || || || |
| Systolic|| || || || || |
| <110||36 (12)||125 (21)||1.0||1.0|| |
| 110–119||73 (24)||198 (33)||1.3||1.2 [0.7–2.1]|| |
| 120–129||116 (37)||204 (34)||2.0||1.9 [1.1–3.1]|| |
|≥130||85 (27)||69 (12)||4.3||3.6 [2.0–6.6]||< 0.0001|
|Diastolic|| || || || || |
| <60||7 (2)||32 (5)||1.0||1.0|| |
| 60–69||67 (22)||177 (30)||1.7||1.6 [0.6–4.1]|| |
| 70–79||144 (46)||246 (41)||2.7||2.1 [0.8–5.1]|| |
| ≥80||93 (30)||141 (24)||3.0||1.8 [0.7–4.6]||0.4|
|Maternal weight (kg)|| || || || || |
| <60||61 (20)||196 (33)||1.0||1.0|| |
| 60–69||108 (35)||237 (40)||1.5||14 [0.9–2.1]|| |
| 70–79||79 (26)||117 (19)||2.2||1.9 [1.2–3.0]|| |
|≥80||57 (19)||50 (8)||3.7||3.0 [1.7–5.3]||< 0.0001|
|Cigarettes per day|| || || || || |
| 0||251 (80)||442 (74)||1.0||1.0|| |
| 1–4||12 (4)||31 (5)||0.7||0.6 [0.3–1.3]|| |
| ≥5||49 (16)||128 (21)||0.7||0.6 [0.4–0.9]||0.01|
|Multiple birth pregnancy|| || || || || |
| No||309 (94)||638 (97)||1.0||1.0|| |
| Yes||14 (6)||12 (3)||2.4||2.0 [0.8–5.0]|| |
Table 3. The association (odds ratio) and 95% confidence intervals (CI) between risk factors recorded at first antenatal visit (before 18 weeks of gestation) and subsequent risk of pre-eclampsia (PE), according to categories of disease severity (mild, moderate, or severe). Values are given as OR (95% CI).
|PE in a previous pregnancy*|| || || |
| Yes||11.6 (4.4–30.5)||28.3 (11.0–72.9)||26.1 (8.5–80.3)|
|Nulliparous|| || || |
| Yes||2.6 (1.6–4.1)||3.2 (2.1–4.9)||4.0 (2.4–6.8)|
|Blood pressure (mmHg)|| || || |
|Systolic|| || || |
| < 120||1.0||1.0||1.0|
| ≥ 120||1.8 (1.1–3.1)||2.4 (1.5–3.9)||1.5 (0.9–2.6)|
|Diastolic|| || || |
| < 80||1.0||1.0||1.0|
| ≥ 80||1.5 (0.9–2.5)||0.7 (0.4–1.3)||1.4 (0.8–2.6)|
|Maternal weight (kg)|| || || |
| ≥ 70||3.7 (2.3–6.0)||1.9 (1.2–3.0)||1.3 (0.8–2.2)|
|Smoking during pregnancy|| || || |
| Yes||0.5 (0.3–0.9)||0.7 (0.4–1.2)||0.5 (0.3–0.9)|
|Multiple birth pregnancies|| || || |
| Yes||1.3 (0.3–5.6)||2.8 (1.0–8.2)||1.2 (0.2–5.8)|
Table 4. The association (odds ratio), with 95% confidence intervals, between risk factors recorded at the first antenatal visit (before week 18 of gestation) and early or late onset of pre-eclampsia (PE). Values are given as OR (95% CI).
|PE in a previous pregnancy†|| || |
| Yes||42.4 (11.9–151.6)||21.3 (9.7–46.8)|
|Nulliparous|| || |
| Yes||3.8 (1.7–8.3)||3.1 (2.9–4.3)|
|Blood pressure (mmHg)|| || |
|Systolic|| || |
| < 120||1.0||1.0|
| ≥ 120||1.7 (0.7–3.9)||2.0 (1.4–2.9)|
|Diastolic|| || |
| < 80||1.0||1.0|
| ≥ 80||1.7 (0.7–4.0)||1.0 (0.7–1.5)|
|Maternal weight (kg)|| || |
| < 70||1.0||1.0|
| ≥ 70||1.7 (0.8–3.7)||2.1 (1.5–3.0)|
|Smoking during pregnancy|| || |
| Yes||0.9 (0.4–2.2)||0.5 (0.4–0.8)|
|Multiple birth pregnancies|| || |
| Yes||3.6 (0.7–18.4)||1.8 (0.7–4.5)|
Information on maternal smoking at 18 weeks of gestation was obtained for 94% of the study population, and 61 cases (20%) and 159 controls (26%) reported smoking. Women who reported smoking had a relative risk of pre-eclampsia of 0.6 (95% CI 0.4–0.9) (data not shown) compared to nonsmokers, and the estimate was not influenced by number of cigarettes per day (Table 2). This result was consistent across categories of disease severity (Table 3), but for early onset pre-eclampsia, there was no association with smoking (Table 4). In the analysis of parous women only (Table 5), there was no association with smoking among women with previous pre-eclampsia, whereas we found a negative association among women without previous pre-eclampsia (OR 0.4; 95% CI 0.2–0.9).
Table 5. The association (odds ratio) between risk factors recorded before week 18 of gestation and subsequent pre-eclampsia, in nulliparous and parous women, by “repeater” status of pre-eclampsia among parous women. Values are given as OR (95% CI).
|Blood pressure (mmHg)|| || || |
|Systolic|| || || |
| < 120||1.0||1.0||1.0|
| ≥ 120||1.2 (0.8–1.9)||2.4 (1.1–5.9)||7.5 (3.5–16.0)|
|Diastolic|| || || |
| < 80||1.0||1.0||1.0|
| ≥ 80||1.3 (0.8–2.2)||1.3 (0.6–2.6)||0.6 (0.3–1.2)|
|Maternal weight (kg)|| || || |
| < 70||1.0||1.0||1.0|
| ≥ 70||2.1 (1.3–3.3)||2.7 (1.4–5.2)||1.7 (0.9–3.1)|
|Smoking during pregnancy|| || || |
| Yes||0.6 (0.4–1.0)||0.8 (0.4–1.6)||0.4 (0.2–0.9)|
|Multiple birth pregnancies|| || || |
| Yes||1.6 (0.4–5.9)||3.1 (0.7–13.6)||2.3 (0.5–11.2)|
High maternal weight was positively associated with the risk of pre-eclampsia (Table 2). Women initially weighing ≥ 80 kg had a relative risk of 3.0 (95% CI 1.7–5.3) compared to women who weighed < 60 kg. The risk increase was confined to mild and moderate preeclampsia, and was not present for severe disease (Table 3). For early onset pre-eclampsia, there was only a weak association with high weight (Table 4). Among parous women with a history of pre-eclampsia, maternal weight was strongly associated with pre-eclampsia in the current pregnancy (Table 5).
Systolic blood pressure at first antenatal visit was positively associated with the risk of pre-eclampsia (Table 2). Women with initial systolic blood pressure ≥ 130 mmHg had a relative risk of 3.6 (95% CI 2.0–6.6) compared to women with systolic blood pressure < 110 mmHg. For diastolic blood pressure at first antenatal visit there was a positive but weak association with the risk of pre-eclampsia (Table 2). Women with initial diastolic blood pressure ≥ 80 mmHg had a relative risk of 1.8 (95% CI 0.7–4.6) compared to having diastolic blood pressure < 60 mmHg.
Multiple birth pregnancies were positively associated with pre-eclampsia (Table 2), but the numbers were few (OR 2.0; 95% CI 0.8–5.0). Nulliparous women were at increased risk of pre-eclampsia compared to parous women (OR 3.6; 95% CI 2.6–5.0). The strength of the association increased slightly with disease severity (Table 3), and was equally strong for both early and late onset pre-eclampsia (Table 4).
The overall risk of pre-eclampsia was 2.5% in this unselected population of nearly 13,000 delivering Norwegian women. The relatively low estimate, compared to previously reported results3,22 can probably be ascribed to a number of circumstances. First, the Norwegian health insurance provides free antenatal and obstetric care to all inhabitants. Thus, the attendance and compliance to a population-based study of pregnancy in Norway most likely will be high. Second, various definitions and classification systems of pre-eclampsia are in international use, and direct comparisons between incidence estimates may not be appropriate5. Since we applied a relatively narrow definition of pre-eclampsia in this study, a low incidence should be expected. Finally, in this study both nulliparous and parous women were included, whereas many previous studies have been restricted to nulliparous women3,4. Since nulliparity is a well accepted risk factor for pre-eclampsia10, studies of nulliparous women can be expected to report a higher frequency of pre-eclampsia than studies which have included all pregnancies in a population.
In accordance with previous findings, we observed an increased risk for pre-eclampsia associated with nulliparity, high blood pressure, high maternal weight and with previous pre-eclampsia4,11,12. The association with nulliparity and with blood pressure displayed no clear preference for any clinical subtype, neither in relation to disease severity nor to gestational timing of initial pre-eclamptic symptoms. However, the risk increase associated with high maternal weight in our study was confined to pre-eclampsia in its mild and moderate form. This observation differs from that reported by Stone et al.11 who found that maternal obesity may also increase the risk of developing severe preeclampsia. The divergent observations may in part be due to different weight range among participants in the studies. The most obese women in the Stone study were heavier (> 10 kg) than women classified with high maternal weight in our study. In addition, a difference in race may also be important. Different gene combinations have been reported in association with pre-eclampsia23,24, and one might suspect that the frequency of pre-eclampsia susceptibility gene combinations varies between races. In accordance with such an assumption, it has been reported that Blacks may have greater susceptibility to pre-eclampsia than Caucasians25. Thus, since a large proportion of Blacks and Hispanics participated in the study performed by Stone et al.11, whereas all participants in our study were Caucasian, comparisons are difficult.
In our study, previous pre-eclampsia increased the risk of pre-eclampsia in the current pregnancy more than twenty-fold. Hospital based case-control studies, which have used other women in maternity wards as controls, have reported a seven to eleven times higher risk10. It is unclear whether the different strength in the association may reflect differences in study design (e.g. selection of controls), or whether confounding with other factors may be important. One such potentially confounding factor could be the change of partner between pregnancies, which may influence pre-eclampsia risk26. However, we have no information on the change of partner in our study, so we are unable to evaluate the contribution of this factor in the analysis.
Overall, we found that smoking was associated with a reduced risk of pre-eclampsia, in accordance with most studies3,12. Since premature delivery is more frequent among smokers27 and pre-eclampsia usually develops towards term, smoking could be spuriously associated with lower risk of pre-eclampsia due to earlier delivery among smokers. In our study, however, smokers and non-smokers had identical gestational age at delivery. Some authors12, but not others3, have reported a doserelated reduction in pre-eclampsia risk mediated by smoking. In our study, no dose-risk relation was found, but very few women reported to smoke more than ten cigarettes per day. It is worth noticing that smoking apparently had no influence on the risk for early onset pre-eclampsia, which is in accordance with one previous report28, and no effect on the risk for preeclampsia among women with repeated disease. This may suggest that these pre-eclampsia subtypes involve pathogenic mechanisms that are too forceful to be influenced by the putative weak vasodilatating or hypovolemic effects of smoking.
Although more than 100 years have passed since preeclampsia was first described, knowledge of its aetiology and pathophysiology is limited. By regarding preeclampsia as one disease entity, the issue may be confused29. The heterogeneous pre-eclamptic syndrome should be considered a result of different pathogenic processes, and some investigators have suggested a distinction between maternal and placental aetiologic factors30. Furthermore, by stratifying cases of pre-eclampsia according to clinical subtypes, important associations may be revealed that otherwise could go unnoticed. Thus, the differential pattern that we observed between various risk factors on the one hand, and clinical subtypes on the other, may support the suggestion that a useful approach to clinical studies of pre-eclampsia is to divide the syndrome into appropriate subgroups. Based on recent observations23,24,29,31,32, relevant subgroups for future studies could be established based on genetic characteristics (maternal, paternal or both) or placental histopathology.
The authors would like to thank the Norwegian Medical Birth Registry for helpful its contribution to this study. Ms M. Bjelland and Ms A. Haugan collected the clinical data, their work is gratefully acknowledged.