American Journal of Epidemiology Copyright 1998 by The Johns Hopkins University School of Hygiene and Public Health All

l rights reserved

Vol. 147, No. 11 Printed in U.S.A.

Comparison of Risk Factors for Preeclampsia and Gestational Hypertension in a Population-based Cohort Study

Helena Salonen Ros,1 Sven Cnattingius,1 and Loren Lipworth 12 The objective of this study was to evaluate and compare risk factor patterns in association with preeclampsia and gestational hypertension. The data were collected from The Swedish Medical Birth Register and include all nulliparas aged 34 years or less who gave birth at the University Hospital of Uppsala, Sweden, during 1987-1993. Of these 10,666 women, 4.4% developed gestational hypertension, and 5.2% developed preeclampsia. The following risk factors were significantly associated with increased risk of preeclampsia: type 1 diabetes (odds ratio = 5.58, 95% confidence interval 2.72-11.43), gestational diabetes (odds ratio = 3.11, 95% confidence interval 1.61-6.00), and twin birth (odds ratio = 4.17,95% confidence interval 2.30-7.55). The positive associations between these variables and the risk of gestational hypertension were weaker and nonsignificant. Compared with underweight women (body mass index < 19.8), obese women (body mass index > 29) had increased risks of both gestational hypertension (odds ratio = 4.85, 95% confidence interval 1.97-11.92) and preeclampsia (odds ratio = 5.19, 95% confidence interval 2.35-11.48). Significantly lower risks of preeclampsia and gestational hypertension were observed for women born outside Nordic countries and in association with maternal smoking and summer birth. The similarities in risk factor patterns may indicate similarities in the biologic mechanisms underlying the two conditions. Am J Epidemiol 1998; 147:1062-70. body mass index; diabetes, gestational; hypertension; pregnancy, multiple; pre-eclampsia; risk factors; smoking

Downloaded from http://aje.oxfordjournals.org/ by guest on December 9, 2012

Pregnancy-induced hypertension, either with proteinuria (preeclampsia) or without proteinuria (gestational hypertension), develops in 2-10 percent of all pregnancies (1-3) and is one of the most important risk factors for maternal and perinatal mortality. Preeclampsia has been associated with intrauterine growth retardation, preterm birth, and maternal and perinatal deaths (4), while the complications associated with gestational hypertension are less frequent and less severe (5, 6). Comparison of the risk factors for preeclampsia and gestational hypertension may provide insight into the etiologic mechanisms related to these conditions. To date, most epidemiologic and etiologic research has focused on preeclampsia, although maternal smoking and leisure time physical activity have been associated with reduced risks of both preeclampsia and gestaReceived for publication April 22, 1997, and accepted for publication December 2, 1997. Abbreviation: ICD-9, International Classification of Diseases, Ninth Revision. 1 Department of Medical Epidemiology, Karolinska Institute, Stockholm, Sweden. 2 Department of Community Medicine, Mount Sinai School of Medicine, New York, NY. Reprint requests to Dr. Helena Salonen Ros, Department of Medical Epidemiology, Karolinska Institute, P.O. Box 281, S-171 77 Stockholm, Sweden.

tional hypertension within the same study base (7, 8). Using the population-based Swedish Medical Birth Register, we have compared, in a well-defined cohort, maternal and pregnancy characteristics associated with the development of preeclampsia or gestational hypertension.
MATERIALS AND METHODS Description of sample

This study is based on data from the Swedish Medical Birth Register, held by the National Board of Health and Welfare. Starting with the first antenatal visit, information is prospectively collected from all hospital births, including demographic data, previous reproductive history, and complications during pregnancy, delivery, and the neonatal period. The Birth Register contains information about more than 99 percent of all births in Sweden (9). All births reported to the Birth Register are validated each year against another population register, using the mother's and infant's unique personal identification number. Similarly, the mother's personal identification number was used to link the Birth Register to the Education Register, held by Statistics, Sweden, to obtain information about the mother's level of formal education (10).

1062

Risk Factors for Pregnancy-induced Hypertension

1063

From 1987 through 1993, the Birth Register recorded births to 28,037 women at the University Hospital of Uppsala, Sweden. This hospital receives all births from women living in Uppsala County (240,000 inhabitants), totaling 3,500-4,000 births per year. For the present study only nulliparous women were included, because the pathophysiology of hypertension during pregnancy differs between parous and nulliparous women (11). In order to focus on pregnancyinduced hypertension and to minimize the chance of including women with essential hypertension, the study was restricted to women aged 34 years or less (TI = 10,666). In Sweden more than 95 percent of the pregnant population attend antenatal care before the 15th gestational week. Prenatal care is standardized and free of charge and includes visits every fourth week up to 24 gestational weeks, then every second week to 36 weeks, and weekly thereafter. At each visit, blood pressure is measured and urine is checked for protein using a dip stick. Ninety percent of the pregnant population have at least nine visits to antenatal care, and the mean number of visits is 12 (12).
Data collection

Maternal age, height, prepregnancy weight, mother' s country of birth, level of education, family situation, smoking, history of infertility, diabetes, single or multiple birth, season of birth, and infant's sex were evaluated as potential risk factors for gestational hypertension or preeclampsia. Information about maternal height, prepregnancy weight, smoking, family situation, and history of infertility was collected at the woman's first visit to antenatal care, when the woman spends about 1 hour with a midwife for an interview and examination. Maternal height was categorized into short (<154 cm), normal (155-174 cm), and tall (>175 cm). Information about maternal smoking is categorized into nonsmoker (i.e., non-daily smoker), moderate smoker (1-9 cigarettes per day), or heavy smoker (at least 10 cigarettes per day). Information about family situation is categorized into living with infant's father or not. Previous infertility problems are expressed as the number of years of involuntary childlessness. Maternal prepregnancy weight has been routinely recorded since 1992. Thus, prepregnancy body mass index (weight (kg)/height (m) ) was calculated for women who gave birth in 1992 or 1993 and was categorized as follows, according to Institute of Medicine (13) recommendations: underweight (body mass index < 19.8); normal (body mass index = 19.8-26.0); overweight (body mass index = 26.1-29.0); and obese (body mass index > 29.0). Information about maternal
Am J Epidemiol Vol. 147, No. 11, 1998

age, mother's country of birth, type of birth (singleton or multiple), infant's sex, and maternal complications during pregnancy was collected when the women were discharged from the hospital. The mother's country of birth was dichotomized into birth within or outside the Nordic countries. The season of the infant's birth was stratified into winter (December through February), spring (March through May), summer (June through August), or fall (September through November). Complications during pregnancy were classified according to the Swedish version of the International Classification of Diseases, Ninth Revision (ICD-9). The diagnoses are noted by a doctor at the time of discharge from the hospital, using a guide sheet, where the definition of the diagnosis is written in clear text beside the ICD-9 code and a check box. Insulindependent diabetes mellitus type 1 and gestational diabetes were coded as ICD-9 codes 648A and 648G, respectively. Hypertensive diseases were coded as follows: gestational hypertension, ICD-9 codes 642D and 642X; and preeclampsia, ICD-9 codes 642E and 642F. Seven women with the diagnosis eclampsia (ICD-9 code 642G) were excluded from the analyses, yielding a final sample size of 10,659 women. The accuracy of the diagnoses was examined by one of us (H. S. R.) using a random sampling procedure to review 25 percent of the individual records in which the diagnosis was gestational hypertension or preeclampsia. Using the information in these records, gestational hypertension and preeclampsia were in this evaluation strictly defined according to the criteria proposed by the National High Blood Pressure Education Program Working Group Report on High Blood Pressure in Pregnancy (14). That is, gestational hypertension was defined as blood pressure of at least 140/90 mmHg, an increase in systolic blood pressure of at least 30 mmHg, or an increase of diastolic blood pressure of at least 15 mmHg (in at least two readings 6 or more hours apart), without proteinuria, occurring after 20 weeks of gestation. Information on normalization of the blood pressure after delivery is not available through the register. Preeclampsia was defined as hypertension accompanied by proteinuria (two urinary protein dip sticks of at least 1+ or 300 mg of protein or more in a 24-hour urine collection). Among 115 pregnancies coded as gestational hypertension, 97 had gestational hypertension according to the notes in the individual records (positive predictive value = 84 percent). Among 148 pregnancies diagnosed with preeclampsia, 137 had preeclampsia according to the individual records (positive predictive value = 93 percent). The data were modeled through logistic regression (15) using the SAS statistical package (SAS Institute,

Downloaded from http://aje.oxfordjournals.org/ by guest on December 9, 2012

1064

Rosetal.

Inc., Cary, North Carolina). Crude odds ratios and 95 percent confidence intervals were calculated to estimate the effects of individual factors on the risk of gestational hypertension or preeclampsia. Multivariate analyses were performed including those factors that significantly influenced the risk of gestational hypertension or preeclampsia in the univariate analyses. The level of significance was set at p < 0.05.
RESULTS

Among the 10,659 nulliparas, 4.4 percent developed gestational hypertension and 5.2 percent developed preeclampsia. Table 1 shows the distribution of women with gestational hypertension or preeclampsia by maternal sociodemographic, lifestyle, and pregnancy characteristics. Rates of preeclampsia were substantially increased for women with type 1 diabetes (21.3 percent) or gestational diabetes (14.3 percent) and for multiple birth pregnancies (18 percent), while the effects of these factors on rates of gestational hypertension were weaker. Table 2 presents univariate odds ratios for the association between maternal and pregnancy characteristics and risk of gestational hypertension or preeclampsia. Type 1 diabetes, gestational diabetes, and twin birth were significantly associated with increased risk of preeclampsia. Women born outside the Nordic countries, maternal smoking, and summer birth were associated with significantly decreased risks of developing both gestational hypertension and preeclampsia. Maternal age, height, education, history of infertility, and infant's sex did not significantly influence the risk of either gestational hypertension or preeclampsia. The multiple logistic regression analyses included only variables that significantly (p < 0.05) influenced risks of gestational hypertension or preeclampsia in the univariate analyses presented in table 2 (i.e., smoking habits, mother's place of birth, type of pregnancy, season of birth, type 1 diabetes, and gestational diabetes). As presented in table 3, the adjusted odds ratios were similar to the crude odds ratios; thus, significant increased risks of preeclampsia were observed for multiple births (odds ratio = 4.17), type 1 diabetes (odds ratio = 5.58), and gestational diabetes (odds ratio = 3.11), while the influence of these factors on the risk of gestational hypertension was of a smaller magnitude and nonsignificant. Compared with Nordic women, women born outside the Nordic countries had significantly decreased risks of both gestational hypertension (odds ratio = 0.25) and preeclampsia (odds ratio = 0.60). Compared with winter births, the risk of preeclampsia was significantly reduced among summer births (odds ratio = 0.68). Information about prepregnancy body mass index

was available only for the 2,848 women who gave birth during or after 1992. When the analyses were restricted to these women, the rates of gestational hypertension and preeclampsia consistently increased with increasing prepregnancy body mass index; thus, underweight women (body mass index < 19.8) had incidence rates of gestational hypertension and preeclampsia of 2.3 percent and 3.1 percent, respectively, while the corresponding rates among obese women (body mass index > 29.0) were 10 percent and 12.9 percent, respectively (table 4). The association did not change when the effect of maternal prepregnancy body mass index was estimated through multiple logistic regression (table 4). Thus, the risks of both gestational hypertension and preeclampsia increased consistently with increasing prepregnancy body mass index, and obese women had odds ratios of 4.85 and 5.19 for gestational hypertension and preeclampsia, respectively. After adjustment was made for prepregnancy body mass index, the directions of the associations between preeclampsia or gestational hypertension and the other studied variables did not change, although a number of the relations became stronger and the widths of the confidence intervals increased (data not shown).
DISCUSSION

Downloaded from http://aje.oxfordjournals.org/ by guest on December 9, 2012

The collective evidence regarding the etiology of preeclampsia and gestational hypertension, although limited and largely focusing on the former, suggests similarities in risk factor patterns and is compatible with the hypothesis that these conditions may have common underlying mechanisms. Nevertheless, the findings have been inconsistent, possibly because of differences in disease definition and in parity status of the women studied. The present study is based on a unique populationbased cohort, and data were collected prospectively during pregnancy. The accuracy of diagnoses registered in the Swedish Medical Birth Register was examined, and the positive predictive values were high. Moreover, the overall incidences of preeclampsia and gestational hypertension in this data set are comparable with those previously reported (1-3, 16, 17), which also confirm the ensurance of the categorization of hypertensive pregnancies in the Swedish Birth Registry. The incidence of essential hypertension increases with age (18) and, in order to minimize the number of misclassified women with essential hypertension, the study was restricted to women aged 34 years or less. In addition, the study was restricted to nulliparas since the incidence of and risk factors for preeclampsia and gestational hypertension are reported to differ between nulliparous and multiparous women (19).
Am J Epidemiol Vol. 147, No. 11, 1998

Risk Factors for Pregnancy-induced Hypertension

1065

TABLE 1. Distribution of pregnancies with gestational hypertension and preedampsia by maternal and pregnancy characteristics among nulliparous women, 34 years or less, Uppsala County, Sweden, 1987-1993
All births (no.) Gestational hypertension No. % No. Preedampsia

%
5.4 5.3 4.8 22.2 3.4 5.2 5.1 5.8 5.4 3.6 3.3 4.8 6.0 5.0 5.2 4.2 3.8 5.2 5.7 4.8 5.7 3.8 3.5

Maternal age (years)

<19

20-29 30-34 Unknown Maternal height (cm) <154 155-174 >175 Unknown Mother's place of birth Nordic countries Outside Nordic countries Unknown Maternal education (years)
<9

588 8,030 2,032 9 236 7,874 800 1,749 9,654 851 154 1,348 4,046 1,407 1,836

20 359

87
0 6 345 32 83 454 11 1 50 183 67 88

3.4 4.5 4.3 0 2.5 4.4 4.0 4.8 4.7 1.3

32 425 98 2
8 406 41 102 521 31 5 64

0.7 3.7 4.5 4.8 4.8 4.2 2.3 4.4

3.4 4.7 4.7 4.3

Downloaded from http://aje.oxfordjournals.org/ by guest on December 9, 2012

10-11
12

243
70 96

13-14
>15

Unknown Cohabiting with infant's father Yes No Unknown Smoking Non-daily smoking 1-9 cigarettes/day >10 cigarettes/day Unknown History of infertility No >1 year Diabetes

1,675 347 9,287

686 686

70 8
411 23 32 376 58 18

71
13 485 39 33 456 52

8,034 1,357 721 547 9,806

853

14 421 45 458
4

2.5 2.6 4.3 5.3 4.4 8.5 5.2 4.4 6.4

4.8 4.1 4.2 4.4 4.5 4.2 4.4

25 24 510 47 536 10 11
543 14 159 144 115 139 276

4.4 5.2 5.5 5.1 21.3 14.3

5.1

No
Typei Gestational diabetes Type of birth Single Multiple Season of birth Winter Spring Summer Fall Infant's sex Boy Girl Total

10,535 47 77 10,581
78 2,601 2,876 2,656 2,526 5,495 5,164 10,659

4 461 5
124 119 111 112 248 218 466

18.0
6.1 5.0 4.3

5.5 5.0 5.4 5.2

281 557

Within this single, well-defined, population-based cohort of nulliparous women, we have found, with a few exceptions, remarkable similarities in risk factor patterns for preedampsia and gestational hypertension, although the magnitude of the observed associations often differed between the two conditions.
Am J Epidemiol Vol. 147, No. 11, 1998

Among women with type 1 diabetes or gestational diabetes or in twin pregnancies, the elevation in risk of preedampsia was strong and statistically significant, whereas the increased risk of gestational hypertension was not significant. Increased prepregnancy body mass index was associated with significant, increased

1066

Ros et al.
TABLE 2. Crude odds ratios (ORs) with 95% confidence intervals (CIs) for gestational hypertension and preeclampsia by univariate analyses of maternal and pregnancy characteristics among nulliparous women, 34 years or less, Uppsala County, Sweden, 1987-1993
Gestational hypertension OR 95% Cl OR

Preeclampsia
95% Cl

Maternal age (years) <19 20-29* 30-34 Maternal height (cm) <154 155-174* >175 Mother's place of birth Nordic countries* Outside Nordic countries Maternal education (years) <9 10-11 12* 13-14 >15 Cohabiting with infant's father Yes* No Smoking Non-daily smoking* 1-9 cigarettes/day >10 cigarettes/day History of infertility No* 1 year Diabetes No* Typei Gestational diabetes Type of birth Single* Multiple Season of birth Winter* Spring Summer Fall Infant's sex Boy* Girl * Reference group.

0.75 1.00 0.95 0.56 1.00 0.91 1.00 0.26 0.77 0.96 1.00 1.01 0.87 1.00 0.75 1.00 0.89 0.51 1.00 1.25 1.00 2.53 1.34 1.00 1.76 1.00 0.85 0.85 0.92 1.00 0.94

0.48-1.20 0.75-1.21 0.25-1.26 0.63-1.32

1.01 1.00 0.90 0.63 1.00 0.99 1.00 0.64 0.94 1.22 1.00 1.05 0.84 1.00 1.08 1.00 0.66 0.58 1.00 1.08 1.00 5.39 3.16 1.00 4.19 1.00 0.80 0.69 0.89 1.00 1.09

0.70-1.47 0.72-1.13 0.31-1.29 0.71-1.38

0.14-0.48 0.53-1.12 0.72-1.28 0.73-1.40 0.61-1.22

0.44-0.92 0.66-1.33 0.93-1.60 0.77-1.45 0.60-1.18

Downloaded from http://aje.oxfordjournals.org/ by guest on December 9, 2012

0.49-1.15

0.77-1.51

0.67-1.18 0.32-0.82

0.49-0.88 0.39-0.88

0.91-1.71

0.79-1.46

0.89-7.16 0.49-3.71

2.64-11.00 1.65-6.03

0.7O-4.41

2.32-7.54

0.66-1.10 0.66-1.11 0.71-1.20

0.64-1.01 0.54-0.88 0.70-1.13

0.78-1.13

0.92-1.29

risks of both conditions by a similar magnitude, whereas risks were decreased among smokers and among women born outside the Nordic countries. Compared with giving birth in winter, giving birth in summer reduced the risks of preeclampsia and gestational hypertension. Data on several factors previously reported to influence the risk of preeclampsia, such as prepregnancy blood pressure (20), period of sexual cohabitation before pregnancy (3), maternal physical activity during pregnancy (8), previous abortions (21), miscarriages

(20), stress (22), and whether the smokers continued or stopped smoking (7), were not available through the register. The pathophysiology of hypertension in preeclampsia is unclear, and hemodynamic measurements in preeclamptic women have not been consistent or conclusive (23). The pathogenesis of preeclampsia involves insufficient trophoblastic invasion (24-26), which may arise from immunologic imbalance between maternal and fetal tissues (27), preserved smooth muscles in the myometrial spiral arteries (25,
Am J Epidemiol Vol. 147, No. 11, 1998

Risk Factors for Pregnancy-induced Hypertension

1067

TABLE 3. Adjusted* odds ratios (ORs) with 95% confidence intervals (CIs) for gestational hypertension and preeclampsia by maternal and pregnancy characteristics among nulliparous women, 34 years or less, Uppsala County, Sweden, 1987-1993
Gestational hypertension Adjusted* OR 95% Cl Adjusted* OR Preeclampsia 95% Cl

Smoking habits Nonsmokert/ 1-9 cigarettes/day >10 cigarettes/day Mother's place of birth Nordic countriest Outside Nordic countries Type of pregnancy Singlet Multiple Season of birth Wintert. Spring Summer Fall Diabetes

1.00 0.86 0.48 1.00 0.25 1.00 1.73 1.00 0.83 0.83 0.91 1.00 2.52 1.46

0.64-1.14 0.29-0.77

1.00 0.64 0.55 1.00 0.60 1.00

0.47-0.85 0.37-0.84

0.14-0.46

0.42-0.88

0.69-4.35

4.17 1.00 0.80 0.68 0.89 1.00 5.58 3.11

2.30-7.55

0.64-1.07 0.64-1.08 0.70-1.18

0.63-1.01 0.53-0.87 0.70-1.13

Downloaded from http://aje.oxfordjournals.org/ by guest on December 9, 2012

Not
Typei Gestational diabetes

0.89-7.18 0.53-4.07

2.72-11.43 1.61-6.00

* Adjustments were made for the other variables included in the table, t Reference group.

TABLE 4. Multiple logistic regression-derived odds ratios (ORs)* with 95% confidence intervals (CIs) for the association between prepregnancy body mass index (BMI) and gestational hypertension or preelcampsia among nulliparous women, 34 years or less, Uppsala County, Sweden, 1992-1993
Maternal pregnancy BMI (weight (kg)/height (m)*) All births (no.) 350 1,720 208 140 430 2,848 Gestational hypertension No. 8 % Preeclampsia 95% Cl

OR 1.00 1.75 1.98 4.85

No.
11 78 19 18 28 154

% 3.1 4.5 9.1 12.9 6.5

OR 1.00 1.51

95% Cl

Underweight (<19.8t) Normal (19.8-26.0) Overweight (26.1-29.0) Obese (>29.0) Missing Total

70 9 14 11 112

2.3 4.1 4.3 10.0 2.6

0.83-3.69 0.75-5.25 1.97-11.92

3.14 5.19

0.79-2.89 1.44-6.83 2.35-11.48

* Adjustments were made for maternal smoking, place of birth, type 1 diabetes, gestational diabetes, and season of birth, t Reference group.

28), and impaired perfusion of the placenta. The placenta in turn may excrete an agent affecting the maternal vascular endothelium (29), platelet function (30), and/or metabolism (31-33). Our findings that multiple birth is an important risk factor for preeclampsia have been reported by others, especially among nulliparous women (34). The association may be due to the large placental size in multiple birth pregnancies, leading to higher maternal exposure to paternal antigen (27) or impaired placental perfusion (29). The pronounced increase in risk of preeclampsia among type 1 diabetics is consistent with that from previous reports (35-37) and may be due to microvascular changes imparing the placental perfusion. LabAm J Epidemiol Vol. 147, No. 11, 1998

oratory data suggest several other possible mechanisms: 1) the lipid metabolism of pregnant diabetics is more altered compared with that of nondiabetics (38), and both increased lipolytic activity and a high ratio of free fatty acids to albumin in sera have been reported in association with preeclampsia (31, 39); 2) high levels of plasma triglycerides among diabetic women could cause endothelial cells to accumulate triglycerides (31), leading in turn to endothelial dysfunction; and 3) compared with nondiabetics, pregnant diabetics have increased urinary excretion of metabolites of thromboxane that acts as a vasoconstrictor and stimulates platelet aggregation (40). The potential for confounding by chronic hypertension of the association between type 1 diabetes and preeclampsia has been

1068

Ros et al.

minimized because, according to the disease classification used in the present study, only women with hypertension occurring after 20 weeks of gestation and normalized blood pressure within 10 days after delivery were included. Our finding that type 1 diabetes is significantly, albeit less strongly, associated with gestational hypertension may reflect a common metabolic component in the pathogenesis of preeclampsia and gestational hypertension. Gestational diabetes was also an important risk factor for preeclampsia, an association that has been reported in some (41-43), but not all (44), previous studies. Laboratory data have shown a relation between preeclampsia and hyperinsulinemia, and it has been suggested that insulin resistance may play an important role in the pathogenesis of preeclampsia (41, 45, 46). In normotensive women with gestational diabetes, abnormal endothelial vasodilator function has been reported (47) and may predispose to preeclampsia. There is no obvious explanation for the observed association in both the present study and in previous reports (20, 48, 49) between increasing prepregnancy body mass index and the risks of gestational hypertension and preeclampsia. Lipid metabolism is influenced by body mass index, and elevated free fatty acids give rise to insulin resistance, possibly through tumor necrosis factor alpha (50), which is expressed to a high extent in fat tissue of obese subjects (51). Insulin resistance is in turn compensated by hyperinsulinemia, leading to hypertension by vasoconstriction (52). Tumor necrosis factor alpha in plasma is reported to be significantly higher in preeclamptic women and may also give rise to endothelial dysfunction (53). It would be of interest to evaluate body fat distribution in association with preeclampsia and gestational hypertension, since persons with a preponderance of abdominal fat deposition tend to be insulin resistant and have several metabolic abnormalities (54). The positive relation between smoking and cardiovascular disease is well established (55), yet smoking has consistently been reported to reduce the risk of preeclampsia (7, 20) and gestational hypertension (7, 56). Sibai (20) found the lowest incidence of preeclampsia among those who quit smoking at the start of pregnancy, whereas a strong protective effect among women who continued to smoke after 20 weeks of pregnancy has also been reported (7). Since the association with both conditions is similar and the fact that when preeclampsia occurs in smokers it is associated with higher morbidity and mortality (57), it is likely that smoking reduces the expression of impaired perfusion of the placenta rather than enhances the placentation. It has been speculated that the prostacy-

clin/thromboxane balance in smokers would be more advantageous (7), but laboratory data from nonpregnant women show that the urinary excretion of thromboxane metabolites is higher in smokers compared with nonsmokers (58). Metabolic alterations in smokers are reported, but the results are incongruent concerning insulin resistance (59, 60), and there is a lack of knowledge about metabolic changes in pregnant smokers. The strong association between the mother's place of birth and the development of gestational hypertension in the present study may be due to biologic differences in the relation between blood pressure and insulin resistance (61). Despite the fact that antenatal care is free of charge and translators are available, women born outside Nordic countries might not attend antenatal care to the same extent as the Nordic born. However, the diagnoses of gestational hypertension and preeclampsia are also based on information provided when the woman is admitted to the delivery ward, when blood pressure and proteinuria are routinely recorded. This would ensure complete case ascertainment, regardless of symptoms during the pregnancy. In future investigations, it would be of interest to study subgroups in the non-Nordic-born population as well as second generation immigrants. In conclusion, the risk factor patterns for preeclampsia and gestational hypertension, evaluated within this single population-based cohort, appear to be similar and differ primarily in the magnitude of the associations, which tend to be stronger for preeclampsia. These findings, in line with previous investigations (7, 8), may indicate common pathways in the biologic mechanisms underlying the two conditions, and the findings also suggest that weight control may be an important and modifiable factor in the prevention of gestational hypertension and preeclampsia.

Downloaded from http://aje.oxfordjournals.org/ by guest on December 9, 2012

ACKNOWLEDGMENTS

Supported by grants from the Swedish Medical Research Council (grant 27X-12276).

REFERENCES 1. Davies M, Czaczkes JW, Sadowsky E, et al. Toxemia of pregnancy in Jerusalem. Isr J Med Sci 1970;6:253-66. 2. Lindmark G, Lindberg B, Hogstedt S. The incidence of hypertensive disease in pregnancy. Acta Obstet Gynecol Scand Suppl 1984;118:29-32. 3. Robillard P-Y, Hulsey T, Pe"rianin J, et al. Association of pregnancy-induced hypertension with duration of sexual cohabitation before conception. Lancet 1994;344:973-5. 4. Chesley L. History and epidemiology of preeclampsiaeclampsia. Clin Obstet Gynecol 1984;27:801-20. Am J Epidemiol Vol. 147, No. 11, 1998

Risk Factors for Pregnancy-induced Hypertension 5. Naeye RL. Maternal blood pressure and fetal growth. Am J Obstet Gynecol 1981;141:780-7. 6. Page EW, Christianson R. Influence of blood pressure changes with and without proteinuria upon outcome of pregnancy. Am J Obstet Gynecol 1976;126:821-33. 7. Marcoux S, Brisson J, Fabia J. The effect of cigarette smoking on the risk of preeclampsia and gestational hypertension. Am J Epidemiol 1989; 130:950-7. 8. Marcoux S, Brisson J, Fabia J. The effect of leisure time physical activity on the risk of preeclampsia and gestational hypertension. J Epidemiol Community Health 1989;43: 147-52. 9. Cnattingius S, Eriksson A, Gunnarskog J, et al. A quality study of a medical birth registry. Scand J Soc Med 1990;18:143-8. 10. Education in the Swedish population. Reports on statisical coordination. (In Swedish). Orebro, Sweden: Statistics Sweden, 1993:6. 11. Eskenazi B, Fenster L, Sidney S, et al. Fetal growth retardation in infants of multiparous and nulliparous women with preeclampsia. Am J Obstet Gynecol 1993;169:1112-18. 12. Aberg A, Lindmark G. Competence and compliance in antenatal care. Experience from Sweden. Int J Technol Assess Health Care 1992;8(suppl l):20-4. 13. Nutrition during pregnancy and lactation. An implementation guide. Washington, DC: Institute of Medicine, National Academy of Sciences, National Academy Press, 1992. 14. National High Blood Pressure Education Program Working Group report on high blood pressure in pregnancy. Am J Obstet Gynecol 1990;163:1691-712. 15. Breslow NE, Day NE, eds. Statistical methods in cancer research. Vol I. The analysis of case-control studies. Lyon: International Agency for Research on Cancer, 1980. (IARC scientific publication no. 32). 16. Hauth JC, Goldenberg RL, Parker CR, et al. Low-dose aspirin therapy to prevent preeclampsia. Am J Obstet Gynecol 1993; 168:1083-91. 17. Sibai BM, Caritis SN, Thorn E, et al. Prevention of preeclampsia with low-dose aspirin in healthy, nulliparous pregnant women. The National Institute of Child Health and Human Development Network of Matemal-Fetal Medicine Units. N Engl J Med 1993;329:1213-18. 18. Davey D, MacGillivray I. The classification and definition of the hypertensive disorders of pregnancy. Am J Obstet Gynecol 1988;158:892-8. 19. Campbell D, Mac Gillivray I, Carr-Hill R. Preeclampsia in second pregnancy. Br J Obstet Gynaecol 1985;92:131-40. 20. Sibai B. Risk factors for preeclampsia in healthy nulliparous women: a prospective multicenter study. Am J Obstet Gynecol 1995;172:642-8. 21. Seidman D, Ever-Hadani P, Stevenson D, et al. The effect of abortion on the incidence of preeclampsia. Eur J Obstet Gynecol Reprod Biol 1989;33:109-14. 22. Landsbergis P, Hatch M. Psychosocial work stress and pregnancy-induced hypertension. Epidemiology 1996;7: 346-51. 23. Easterling T, Benedetti T, Carlson K, et al. The effect of maternal hemodynamics on fetal growth in hypertensive pregnancies. Am J Obstet Gynecol 1991;165:902-6. 24. de Groot C, O'Brien T, Taylor R. Biochemical evidence of impaired trophoblastic invasion of decidual stroma in women destined to have preeclampsia. Am J Obstet Gynecol 1996; 175:24-9. 25. Meekins JW, Pijnenborg R, Hanssens M, et al. A study of placental bed spiral arteries and trophoblast invasion in normal and severe pre-eclamptic pregnancies. Br J Obstet Gynaecol 1994;101:669-74. 26. Pijnenborg R, Luyten C, Vercruysse L, et al. Attachment and differentiation in vitro of trophoblasts from normal and preeclamptic human placentas. Am J Obstet Gynecol 1996;175: 30-6. 27. Sibai B. Immunologic aspects of preeclampsia. Clin Obstet Gynecol 1991;34:27-34. Am J Epidemiol Vol. 147, No. 11, 1998

1069

28. Khong Y, Sawyer H, Heyret A. An immunologic study of endothelialization of uteroplacental vessels in human pregnancyevidence that endothelium is focally disrupted by trophoblasts in preeclampsia. Am J Obstet Gynecol 1992; 167: 751-6. 29. Roberts J, Redman C. Pre-eclampsia: more than pregnancyinduced hypertension. Lancet 1993;341:1447-51. 30. Friedman S. Preeclampsia: a review of the role of the prostaglandins. Obstet Gynecol 1988;77:122-37. 31. Endresen M, Lorentzen B, Henriksen T. Increased lipolytic activity and high ratio of free fatty acids to abumin in sera from women with preeclampsia leads to triglyceride accumulation in cultured endothelial cells. Am J Obstet Gynecol 1992;167:440-7. 32. Gratac6s E, Casals E, Sanllehy C, et al. Variation in lipid levels during pregnancy in women with different types of hypertension. Acta Obstet Gynecol Scand 1996;75:896-901. 33. Kaaja R, Tikkanen M, Viinikka L, et al. Serum lipoproteins, insulin, and urinary prostanoid metabolites in normal and hypertensive pregnant women. Obstet Gynecol 1995;85:353-6. 34. Coonrod D, Hickok D, Zhu K, et al. Risk factors for preeclampsia in twin pregnancies: a population-based cohort study. Obstet Gynecol 1995;85:645-50. 35. Garner P, D'Alton M, Dudley D, et al. Preeclampsia in diabetic pregnancies. Am J Obstet Gynecol 1990;163:505-8. 36. Hanson U, Persson B. Outcome of pregnancies complicated by type-1 insulin-dependent diabeties in Sweden: acute pregnancy complications, neonatal mortality and morbidity. Am J Perinatol 1993;10:330-3. 37. Siddiqi T, Rosenn B, Mimouni F, et al. Hypertension during pregnancy in insulin dependent diabetic women. Obstet Gynecol 1991;77:514-19. 38. Knopp R, Warth M, Carrol C. Lipid metabolism in pregnancy. I. Changes in lipoprotein triglyceride and cholesterol in normal pregnancy and the effects of diabetes mellitus. J Reprod Med 1973;10:95-106. 39. Potter JM, Nestel PJ. The hyperlipidemia of pregnancy in normal and complicated pregnancies. Am J Obstet Gynecol 1979;133:165-70. 40. Van Assche A, Spitz B, Hanssens M, et al. Increased thromboxane formation in diabetic pregnancy as a possible contributor to preeclampsia. Am J Obstet Gynecol 1993;168:84-7. 41. Abundis E, Ortiz M, Galvan A, et al. Hyperinsulinemia in glucose-tolerant women with preeclampsia. Am J Hypertens 1996;9:610-14. 42. Goldman M, Kitzmiller JL, Abrams B, et al. Obstetric complications with GDM. Effects of maternal weight. Diabetes 1991;40(suppl 2):79-82. 43. Suhonen L, Teramo K. Hypertension and preeclampsia in women with gestational glucose intolerance. Acta Obstet Gynecol Scand 1993;72:269-72. 44. Shaffir J, Lockwood C, Lapinski R, et al. Incidence of pregnancy-induced hypertension among gestational diabetics. Am J Perinatol 1995;12:252-4. 45. Fuh M, Yin C, Pei D, et al. Resistance to insulin-mediated glucose uptake and hyperinsulinemia in women who had preeclampsia during pregnancy. Am J Hypertens 1995;8: 768-71. 46. Sowers JR, Standley PR, Jacober S, et al. Postpartum abnormalities of carbohydrate and cellular calcium metabolism in pregnancy induced hypertension. Am J Hypertens 1993;6: 302-7. 47. Knock G, Me Carthy A, Lowy C, et al. Association of gestational diabetes with abnormal maternal vascular endothelial function. Br J Obstet Gynaecol 1997;104:229-34. 48. Eskenazi B, Fenster L, Sidney S. A multivariate analysis of risk factors for preeclampsia. JAMA 1991;266:237-41. 49. Stone J, Lockwood C, Berkowitz G, et al. Risk factors for severe preeclampsia. Obstet Gynecol 1994;83:357-61. 50. Hotamisligil G, Peraldi P, Budavari A, et al. IRS-1-mediated inhibition of insulin receptor tyrosine kinase activity in TNF-a and obesity-induced insulin resistance. Science 1996;271:665-8.

Downloaded from http://aje.oxfordjournals.org/ by guest on December 9, 2012

1070

Rosetal. incidence but increases the rates of perinatal mortality, abruptio placentae, and intrauterine growth restriction. Am J Obstet Gynecol 1997;177:156-61. Rangemark C, Wennmalm A. Smoke-derived nitric oxide and vascular prostacyclin are unable to counteract the platelet effect of increased thromboxane formation in healthy female smokers. Clin Physiol 1996;16:301-15. Langhoff-Roos J, Wibell L, Gebre-Medin M, et al. Effects of smoking on maternal glucose metabolism. Gynecol Obstet Invest 1993,36:8-11. Nilsson P, Lind L, Pollare T, et al. Increased level of hemoglobin A lc , but not impaired insulin sensitivity, found in hypertensive and normotensive smokers. Metabolism 1995; 44:557-61. Saad M, Lillioja S, Nyomba B, et al. Racial differences in the relation between blood pressure and insulin resistance. N Engl JMed 1991;324:733-9.

51. Hotamisligil G, Arner P, Caro J, et al. Increased adipose tissue expression of tumor necrosis factor-a in human obesity and insulin resistance. J Clin Invest 1995;95:2409-15. 52. Axelrod L. Insulin, prostaglandins, and the pathogenesis of hypertension. Diabetes 1991;40:1223-7. 53. Vince G, Starkey P, Austgulen R, et al. lnterleukin-6, tumor necrosis factor, and soluble tumor necrosis factor receptors in women with preeclampsia. Br J Obstet Gynaecol 1995;102: 20-5. 54. Flack J, Sowers J. Epidemiologic and clinical aspects of insulin resistance and hyperinsulinemia. Am J Med 1991; 91(suppl 1A):11-21. 55. Krupski WC. The peripheral vascular consequences of smoking. Ann Vase Surg 1991;5:291-304. 56. Misra D, Kiely J. The effect of smoking on the risk of gestational hypertension. Early Hum Dev 1995;40:95-107. 57. Cnattingius S, Mills JL, Yuen J, et al. The paradoxical effect of smoking in preeclamptic pregnancies: smoking reduces the

58.

59. 60.

61.

Downloaded from http://aje.oxfordjournals.org/ by guest on December 9, 2012

Am J Epidemiol

Vol. 147, No. 11, 1998