Fetal growth restriction (FGR) is the end point of a number of pregnancy-associated conditions, and the mechanisms that lead to it differ.1 Fetal growth can be restricted by preeclampsia,2 but most infants born to women with preeclampsia weigh appropriate for their gestation.2 Preeclampsia should probably be regarded as a syndrome of heterogeneous origin.2 Shallow trophoblast invasion of decidual arteries can precipitate preeclampsia,3 reduce placental perfusion, and cause insufficient transport of nutrients. Placental morphologic changes vary substantially in preeclampsia,2,4 and it has been hypothesized that FGR might depend on abnormal placental development.2 In cases in which maternal factors (genetic, metabolic, hemodynamic) are dominant, placental perfusion is not necessarily affected and has little impact on fetal growth.2
Clinical manifestations of preeclampsia vary by gestational age at onset (early or late) and by severity of symptoms (mild, moderate, severe). Placental disease has been reported as a consistent characteristic of early preeclampsia,4,5 and that corresponds to the serious reduction in birth size associated with those cases.6–8 Reduced birth size has also been seen after clinically severe preeclampsia with later onset,9,10 but other studies did not show differences in growth between mild and severe preeclampsia.11
Among maternal factors, growth restriction caused by smoking during pregnancy is an established risk factor.9,12 A synergistic effect has been suggested when smoking is combined with preeclampsia, causing lower birth weight than expected by adding their separate effects.9 However, the results of others do not support a synergy between smoking and preeclampsia on fetal growth.13
In this population-based study, we examined the association between different clinical manifestations of preeclampsia and fetal growth and explored whether a relationship between preeclampsia and fetal growth could be modified by maternal factors, particularly smoking.
Materials and Methods
The study was done between January 1993 and December 1995 at the Central Hospital in Rogaland County, Norway. The birthing clinic at this hospital exclusively serves a region of approximately 239,000 inhabitants, and there were 12,804 deliveries during the study period. The study was considered and approved by the Regional Committee for Ethics in Medical Research.
Since 1967, the Norwegian Medical Birth Registry has used standardized forms to record information on all deliveries.14 We searched the records of the Birth Registry to identify women with preeclampsia who gave birth at Rogaland Central Hospital during the study period and found approximately 1300 cases with clinical characteristics possibly indicative of preeclampsia. For each potential case we verified and supplemented that information with detailed clinical information from hospital records. After reviewing all relevant records, we found that 323 women fulfilled the diagnostic criteria for preeclampsia. After that, the Medical Birth Registry selected two separate groups of women without preeclampsia who gave birth at the hospital during the same period. One group consisted of the first women who gave birth at the birthing clinic after the women with preeclampsia. The other group was randomly selected by computer among all other births at the hospital, but frequency matched by mother's age to avoid confounding between effect of preeclampsia and maternal age.
Using each control group separately in the analysis yielded almost identical results, so we decided to pool the two groups to increase statistical precision. The results presented are based on those pooled analyses. We excluded women with twin pregnancies and women with unknown gestational ages from analysis, leaving 307 live singleton infants born after preeclamptic pregnancies and 619 controls.
We used a reported definition of preeclampsia,15 ie, persistent diastolic blood pressure (BP) of at least 90 mmHg had to develop after 20 weeks' gestation and it had to increase by at least 25 mmHg. Nineteen women (four with histories of hypertension) had diastolic BP of 90 mmHg at baseline, and they were included as cases of preeclampsia because their diastolic BP increased further by at least 15 mmHg. Proteinuria also had to be present for preeclampsia; when cutoff was defined as 0.3 mg/L (semiquantitative dipstick 1+) in at least one urine sample after 20 weeks' gestation, without simultaneous urinary infection. Twelve women with no histories of hypertension had no registered baseline BP, but had diastolic pressures of 105 mmHg or higher after 20 weeks' gestation (with proteinuria). They were included as cases of preeclampsia.
Preeclampsia was categorized as mild, moderate, or severe.16 Mild preeclampsia was defined as diastolic BP increase of at least 25 mmHg and proteinuria of 1+ on semiquantitative dipstick; moderate preeclampsia as an increase in diastolic BP of at least 25 mmHg and proteinuria of 2+ on semiquantitative dipstick; and severe preeclampsia as diastolic BP increased to at least 110 mmHg and proteinuria of 3+ on semiquantitative dipstick, or at least 500 mg/24 hours. Six cases of eclampsia and 16 cases with indications of hemolysis, elevated liver enzymes, low platelets (HELLP) syndrome were classified as severe preeclampsia. Pregnancy termination before or at 32 weeks' gestation was treated as a proxy variable for early-onset preeclampsia.
The primary outcome of this study was expressed as the ratio between observed and the expected birth weight (birth weight ratio),17 in which the expected birth weight was adjusted for sex and gestational age at birth. Gestational age was calculated exclusively from routine ultrasonographic measurements of biparietal diameter at 18 weeks' gestation according to Norwegian standard curves.18 Weight curves estimated from ultrasonographic measurements in a population of healthy pregnant Swedish women were used to determine expected birth weights for sex and gestational age.19
A small for gestational age (SGA) infant was defined as having a birth weight two standard deviations or more below the expected birth weight, which corresponds to more than 24% lower birth weight than expected (birth weight ratio less than 0.76), or an approximately 840-g reduction in birth weight for a term infant.
In Norway, antenatal care is free, and most women (close to 100%) attend their first antenatal doctors' visits around 12 weeks' gestation. Clinical information is recorded on standardized forms, and the antenatal maternal data analyzed in this study were based on that information. Few women reported smoking more than ten cigarettes per day, so participants were dichotomized as smokers or nonsmokers. Maternal weight was measured at first antenatal visit, classified as prepregnancy weight, divided into the following three categories: under 60, 60–79, and at least 80 kg. In the analysis of the risk of SGA, maternal weight was dichotomized at 70 kg, and parity as nulliparous or parous.
Student t test was used for comparison of continuous variables between groups. To compare proportions, as indicated by categoric variables, we used χ2 test. Birth weight ratios of infants whose mothers had preeclampsia were calculated and compared with weight ratios of control infants. That comparison was stratified according to mother's parity, maternal smoking, and the three categories of maternal weight at first antenatal visit, and covariates were included in a multiple linear regression analysis to control for potential confounding. We estimated the odds ratio (OR) for SGA as a measure of relative risk (RR) between infants whose mothers had preeclampsia and control infants and used unconditional logistic regression to adjust for potentially confounding factors in a multivariate analysis.20 We further explored whether maternal factors (parity, smoking and prepregnancy weight) could modify associations between subgroups of preeclampsia and birth size and tested possible interactions in multivariate models (linear regression for birth weight ratio and logistic regression for SGA). Precision of the estimates of effect (birth weight ratio and OR) were estimated with 95% confidence intervals (CI). All statistical analyses were calculated using the Statistical Package for the Social Sciences (SPSS) (SPSS Inc., Chicago, IL).
Birth status of infants whose mothers had preeclampsia and controls is shown in Table 1. The mean birth weight was 5% (95% CI 3%, 6%) lower than expected in the preeclampsia group (Table 2). That corresponded to an approximately 175-g lower birth weight than expected for a term infant. Stratified analyses (Table 2) showed that the birth weight was 10% (95% CI 6%, 14%) lower than expected in newborns whose mothers had preeclampsia and reported smoking, compared with 3% (95% CI 1%, 5%) for mothers who had preeclampsia but did not smoke. Among control infants, newborns of smokers also weighed less than expected (4%, 95% CI 3%, 6%). There was no association between maternal baseline BP and birth weight in the preeclampsia group or among control infants (data not shown). We tested for statistically significant interactions between preeclampsia and all the maternal factors listed in Table 2, but found none (P ≥ .10). We also explored possible statistical interactions between smoking and clinical subtypes of preeclampsia, but there was no interaction with any subtypes (data not shown).
The risk of having an SGA infant (Table 3) was four times higher (RR equals; 4.2, 95% CI 2.2, 8.0) in women with preeclampsia than controls. In stratified analyses, we evaluated the association between preeclampsia and risk of SGA for different categories of maternal smoking, parity, and prepregnant weight. The results showed that among paras, the RR of having an SGA infant was 7.9 (95% CI 2.8, 22.2) in women with preeclampsia compared with controls. Among nulliparas the RR was 2.8 (95% CI 1.2, 5.9). The test of statistical interaction between preeclampsia and parity was not significant (P = .08). Women with recurrent preeclampsia were at particularly high risk (RR = 12.3, 95% CI 3.9, 39.2) of having SGA infants compared with controls.
Table 4 shows that birth size was lower with increasing severity of preeclampsia (P trend < .01) and that risk of having an SGA infant increased with disease severity (P trend = .05). For severe preeclampsia, birth weight was 12% (95% CI 9%, 15%) lower than expected, but after mild preeclampsia, birth weight did not differ from the expected weight. The proportion of SGA infants born after severe preeclampsia was 21% (95% CI 12%, 29%) compared with 6% (95% CI 1%, 10%) after mild preeclampsia.
Early-onset preeclampsia (Table 5) was strongly associated with low birth weight, 23% (95% CI 18%, 29%) lower than expected, and in early-onset preeclampsia, the frequency of SGA infants was 53% (95% CI 36%, 70%). We also distinguished between severe preeclampsia with late onset and early onset (Table 6), and the results showed that severe preeclampsia relatively late in pregnancy also was related to lower than expected birth weight (9%, 95% CI 6%, 12%).
Neonates whose mothers had preeclampsia weighed less than infants born after normotensive pregnancies, and their risk of being born SGA was fourfold higher. Thus, our results agreed with those that reported reduced fetal growth in preeclampsia.9,12,21 Weight reduction differed strongly between clinical subgroups and was mainly confined to infants whose mothers had early-onset or severe preeclampsia. The most serious growth restriction (23% lower than expected) was in the early-onset group, and more than half of those newborns were SGA.
Ness and Roberts2 hypothesized that preeclampsia restricts fetal growth when it is caused by placental abnormalities, which result in reduced nutrient supply to the fetus. The serious FGR that accompanies early-onset preeclampsia and the abundant uteroplacental vascular lesions in placental tissues associated with early-onset preeclampsia4 fit well with that hypothesis. Abnormal observations are less frequent in placental tissue from preeclamptic deliveries at term,5 and mild and moderate preeclampsia appear to have only negligible effects on birth weight, as reported by others9 and supported by our data.
Although the association between early-onset preeclampsia and SGA is well established, results have varied substantially, ranging between 18% and 80%.6–8 However, divergent definitions of early-onset preeclampsia were used, which might account for some variation. To some extent, early-onset and severe preeclampsia are overlapping categories. In our early-onset group, two thirds of women were classified as having severe symptoms. Although less pronounced, there was also lower birth size related to severe preeclampsia with late onset (9% lower than expected), which is in accordance with previous reports. Cnattingius et al9 found a substantially higher risk of SGA in pregnancies with severe than mild preeclampsia. However, a recent Chinese study reported no difference in risk of SGA infants for women with mild and severe preeclampsia.11
Ultrasound measurements have become the standard method for pregnancy dating in Scandinavia because ultrasound might predict delivery date more precisely than last menstrual period.22,23 Gestational ages in the present study were determined by ultrasound. We used fetal growth curves based on ultrasound measurements to estimate expected birth weights for two reasons.19 First, by using identical methods for pregnancy dating and evaluation of fetal growth,24 precision and validity of estimated deviations from expected growth (birth weight ratio) can be improved. Second, postnatal measurements to construct birth weight standards in pre-term infants have been criticized because the underlying pathogenesis of preterm parturition might restrict fetal growth and cause a lower birth weight than indicated by gestational age.24,25 Therefore, FGR might be underestimated in premature newborns,12 and expected birth weights derived from weight curves based on ultrasound will be slightly higher than those from postnatal weight standards.19,26 In control infants born between 231 and 302 days' gestation, birth weights were practically identical to those expected from ultrasound-based weight curves for them. That might be reassuring for the validity of the method that we used for that range of gestational age. At lower gestational ages, however, we are not provided with similar healthy control infants, so we cannot exclude that growth restriction in 37 infants born before 231 days' gestation was overestimated by ultrasound-based weight curves. Given the magnitude of the growth restriction related to early-onset preeclampsia, it seems unlikely that more than a fraction can be ascribed to possible bias in the estimates.
In control infants, we found that maternal smoking was related to a reduction in birth weight of 4% and that birth weight after preeclampsia in nonsmokers was 3% lower than expected. Birth weight after preeclampsia in smokers was reduced by 10%, which indicated an additive statistical effect of preeclampsia and smoking on birth size. That finding might be an argument against suggested synergy between smoking and preeclampsia.9,10,27 Our results might suggest that smoking influences fetal growth by mechanisms that are independent of, and not interacting with, mechanisms in preeclampsia that also restrict growth.
Among parous women, the risk of SGA associated with preeclampsia was substantially higher than the risk for nulliparas. Similar observations were reported by Eskenazi et al.21 We found that women with preeclampsia who had it in previous pregnancies had dramatically higher risk of delivering an SGA infant. The distribution of clinical subtypes (mild, moderate, severe, and late versus early onset) (data not shown) did not differ between nulliparas and paras with preeclampsia, regardless of whether they had it before. The effect on fetal growth of repeated preeclampsia was also present in cases in which clinical severity was only moderate. It remains unknown whether paras with preeclampsia have a separate disease in origin or pathogenesis from preeclampsia in nulliparas. The serious growth restriction associated with recurrent preeclampsia suggests that future studies should focus on placental histopathology associated with various clinical subgroups.
1. Arnholdt H, Meisel F, Fandrey K, Lohrs U. Proliferation of villous trophoblast of the human placenta in normal and abnormal pregnancies. Virchows Arch B Cell Pathol Incl Mol Pathol 1991;60:365–72.
2. Ness RB, Roberts JM. Heterogeneous causes constituting the single syndrome of preeclampsia: A hypothesis and its implications. Am J Obstet Gynecol 1996;175:1365–70.
3. Roberts JM, Redman CW. Preeclampsia: More than pregnancy-induced hypertension. Lancet 1993;341:1447–51.
4. Ghidini A, Salafia CM, Pezzullo JC. Placental vascular lesions and likelihood of diagnosis of preeclampsia. Obstet Gynecol 1997;90:542–5.
5. Teasdale F. Histomorphometry of the human placenta in maternal preeclampsia. Am J Obstet Gynecol 1985;152:25–31.
6. Long PA, Abell DA, Beischer NA. Fetal growth retardation and preeclampsia. Br J Obstet Gynaecol 1980;87:13–8.
7. Brazy JE, Grimm JK, Little VA. Neonatal manifestations of severe maternal hypertension occurring before the thirty-sixth week of pregnancy. J Pediatr 1982;100:265–71.
8. Moore MP, Redman CW. Case-control study of severe preeclampsia of early onset. BMJ 1983;287:580–3.
9. Cnattingius S, Mills JL, Yuen J, Eriksson O, Salonen H. The paradoxical effect of smoking in preeclamptic pregnancies: Smoking reduces the incidence but increases the rates of perinatal mortality, abruptio placentae, and intrauterine growth restriction. Am J Obstet Gynecol 1997;177:156–61.
10. Duffus GM, MacGillivray I. The incidence of preeclamptic toxaemia in smokers and non-smokers. Lancet 1968;1:994–5.
11. Xiong X, Mayes D, Demianczuk N, Olson DM, Davidge ST, Newburn CC, et al. Impact of pregnancy-induced hypertension on fetal growth. Am J Obstet Gynecol 1999;180:207–13.
12. Spinillo A, Capuzzo E, Piazzi G, Nicola S, Colonna L, Iasci A. Maternal high-risk factors and severity of growth deficit in small for gestational age infants. Early Hum Dev 1994;38:35–43.
13. 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.
14. Lie RT, Rasmussen S, Brunborg H, Gjessing HK, Lie NE, Irgens LM. Fetal and maternal contributions to risk of preeclampsia: Population based study. BMJ 1998;316:1343–7.
15. CLASP: A randomised trial of low-dose aspirin for the prevention and treatment of preeclampsia among 9364 pregnant women. CLASP (Collaborative Low-dose Aspirin Study in Pregnancy) Collaborative Group. Lancet 1994;343:619–29.
16. Redman CWE. Hypertension in pregnancy. New York: Perinatology Press, 1987.
17. Kramer MS, McLean FH, Olivier M, Willis DM, Usher RH. Body proportionality and head and length ‘sparing’ in growth-retarded neonates: A critical reappraisal. Pediatrics 1989;84:717–23.
18. Eik-Nes SH, Grøttum P, Jørgensen NP, Løkvik B. Normal range curves for BPD and MAD. Drammen, Norway: Scan-Med A/S, 1983.
19. Marsal K, Persson PH, Larsen T, Lilja H, Selbing A, Sultan B. Intrauterine growth curves based on ultrasonically estimated foetal weights. Acta Paediatr 1996;85:843–8.
20. Kleinbaum D. Logistic regression. A self-learning text. New York: Springer-Verlag, 1998.
21. Eskenazi B, Fenster L, Sidney S, Elkin EP. Fetal growth retardation in infants of multiparous and nulliparous women with preeclampsia. Am J Obstet Gynecol 1993;169:1112–8.
22. Tunon K, Eik-Nes SH, Grøttum P. A comparison between ultrasound and a reliable last menstrual period as predictors of the day of delivery in 15,000 examinations. Ultrasound Obstet Gynecol 1996;8:178–85.
23. Gardosi J, Geirsson RT. Routine ultrasound is the method of choice for dating pregnancy. Br J Obstet Gynaecol 1998;105:933–6.
24. Pollack RN, Divon MY. Intrauterine growth retardation: Definition, classification, and etiology. Clin Obstet Gynecol 1992;35:99–107.
25. Seeds JW. Impaired fetal growth: Definition and clinical diagnosis. Obstet Gynecol 1984;64:303–10.
26. Bernstein IM, Mohs G, Rucquoi M, Badger GJ. Case for hybrid “fetal growth curves”: A population-based estimation of normal fetal size across gestational age. J Matern Fetal Med 1996;5:124–7.
© 2000 The American College of Obstetricians and Gynecologists
27. Salafia C, Shiveric K. Cigarette smoking and pregnancy II: Vascular effects. Placenta 1999;20:273–9.