Analysis of mid-trimester corticotrophin-releasing hormone and {alpha}-fetoprotein concentrations for predicting pre-eclampsia

T.N. Leung1,6, Tony K.H. Chung1, Gemma Madsen2, C.W.K. Lam3, Peggo K.W. Lam4, William A.W. Walters5 and Roger Smith2

1 Department of Obstetrics and Gynaecology, The Chinese University of Hong Kong, Hong Kong, 2 The Mothers and Babies Research Centre, University of Newcastle, John Hunter Hospital, Australia, 3 Department of Chemical Pathology, 4 Centre for Clinical Trials and Epidemiological Research, The Chinese University of Hong Kong, Hong Kong, 5 Discipline of Reproductive Medicine and Department of Obstetrics and Gynaecology, University of Newcastle, John Hunter Hospital, Australia


    Abstract
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Reference
 
The aims of this study were firstly to examine if corticotrophin-releasing hormone (CRH) concentrations in maternal plasma were significantly elevated in Chinese pregnancies complicated by pre-eclampsia, secondly to assess if this elevation could be detected in the mid-trimester before onset of clinical signs of the disease, and thirdly to evaluate the performance of using maternal CRH and/or {alpha}-fetoprotein (AFP) concentrations in the mid-trimester for prediction of pre-eclampsia. The first part of this study was tested in a cohort of 39 subjects. The CRH concentrations were significantly elevated in pregnant women complicated by pre-eclampsia. The second and third parts of the study involved a different cohort of 1021 subjects. Both CRH and AFP concentrations in the mid-trimester were significantly elevated in those who subsequently developed pre-eclampsia. However, when used for prediction of pre-eclampsia, neither the CRH nor AFP concentrations alone in the mid-trimester had strong predictive value. Although the combination of both tests improved the detection rate compared to the use of CRH alone, the small increase in the likelihood ratio from 1.9 to 2.6 did not suggest that the combination would be of great clinical value.

Key words: alpha-fetoprotein/corticotrophin releasing hormone/pre-eclampsia/prediction


    Introduction
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Reference
 
Pre-eclampsia is a multi-system disorder specific to pregnant women. It remains one of the most important causes of maternal and fetal mortality and morbidity in developed countries (Redman, 1991Go). The pathogenesis of this condition is not fully understood but much evidence suggests that one of the major underlying pathological changes occurs in the placental bed (Robertson et al., 1967Go; Roberts and Redman, 1993Go). There is failure of, or incomplete trophoblastic invasion of, the spiral arteries resulting in placental ischaemia (Roberts and Redman, 1993Go; Lyail and Greer, 1994Go). Subsequently, the poorly perfused fetoplacental unit may release one or more factors which damage maternal vascular endothelial cells leading to the multisystem dysfunction (Roberts and Redman, 1993Go; Lyail and Greer, 1994Go). The normal process of trophoblastic invasion is complete by 20 weeks of gestation. Hence, the initiating pathology should exist prior to this stage of pregnancy, long before the onset of the clinical syndrome. Therefore, if poor placentation is the major predisposing factor for pre-eclampsia, it is theoretically possible to identify those subjects at risk of developing the condition by screening at early gestation. The early identification of a high-risk group would allow assessment of the effectiveness of prophylactic measures in this condition.

Corticotrophin-releasing hormone (CRH) is synthesized by the placenta during human pregnancy and released into both maternal and fetal circulations. Abnormal elevation of CRH concentrations has been reported in pregnancies complicated by pregnancy induced hypertension (PIH) and pre-eclampsia (PE) (Wolfe et al., 1988Go; Laatikainen et al., 1991Go; Emanuel et al., 1994Go; Perkins et al., 1995Go; Liapi et al., 1996Go). A rise in CRH concentrations has also been demonstrated 11 weeks before the development of signs and symptoms of the disease (Wolfe et al., 1988Go) although Petraglia et al. (1996) showed that changes in CRH and CRH binding protein concentrations did not precede the onset of clinical disease. In most of these studies, the `hypertensive' group consisted of a combination of pregnant women with PIH (without proteinuria) and PE (with proteinuria) (Wolfe et al., 1988Go; Laatikainen et al., 1991Go; Emanuel et al., 1994Go; Liapi et al., 1996Go; Petraglia et al., 1996Go). While PE is a more stringent diagnosis specific to pregnancy, PIH may include a group of women with transient hypertension and its inclusion may lead to a heterogeneous cohort (Roberts and Redman, 1993Go). In the current study, data were collected on CRH concentrations from over 1500 Chinese pregnant subjects between 15–42 weeks of gestation. Firstly the difference in CRH concentrations in the third trimester between normal pregnant subjects and those with established PE was studied. Then the CRH data from the mid-trimester samples were analysed to test if concentrations were elevated in those who subsequently developed PE. The clinical utility of plasma CRH for use in the prediction of PE was then assessed.

Alpha-fetoprotein (AFP) is a valued screening test for both neural tube defects and biochemical screening for Down's syndrome. Pre-eclampsia has also been associated with an elevation of AFP in maternal serum (Moore and Redman, 1983Go; Walters et al., 1985Go). Using a threshold value of two multiples of the median (MoM), elevated AFP in the mid-trimester has been shown to be associated with a 2.3- to 3.8-fold increased risk of developing pre-eclampsia (Milunsky et al., 1989Go; Williams et al., 1992Go). The mid-trimester AFP concentrations for over 1000 subjects in this cohort were available for analysis. The performance using AFP alone or in combination with CRH for prediction of pre-eclampsia was also examined.


    Materials and methods
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Reference
 
Study population
The study population comprised 1523 Chinese pregnant women attending the antenatal clinic or antenatal ward of a tertiary hospital in Hong Kong. They were recruited at random by a research assistant between April 1996 and October 1997. Ethical approval for the study was obtained from the institutional Human Research Ethics Committee of the Chinese University of Hong Kong. Only singleton pregnancies were included. All subjects' gestational ages were confirmed by early ultrasound examination. The mid-trimester CRH data of 1014 subjects in this cohort and their performance in prediction of preterm delivery have been described in detail in a previous publication (Leung et al., 1999Go).

Sampling details
Ten ml of venous blood were collected by venepuncture of the antecubital vein of each subject between 15 and 42 weeks gestation and placed immediately into ice-cold lithium heparin tubes. After centrifugation at 3000 g for 20 min at 4°C, the plasma was collected in duplicate aliquots and stored at –70°C until assay for CRH. A total of 1021 subjects recruited between 15–22 weeks also had AFP concentrations assayed.

CRH assay
Frozen plasma samples were transported in dry ice to The Mothers and Babies Research Centre, University of Newcastle, Australia for measurement of CRH by radioimmunoassay as previously described (Leung et al., 1999Go). The mean extraction recovery for the radioimmunoassay was 103% and the intra- and interassay coefficients of variation were 7% and 10% respectively. The threshold for detection was 1.35 pmol/l. The investigators responsible for the laboratory assay were blinded to the obstetric outcome.

AFP assay
AFP concentrations were measured by microparticle enzyme immunoassay (Imx analyzer®; Abbott Laboratories, Abbott Park, IL, USA) in the Department of Chemical Pathology, Prince of Wales Hospital, Shatin, Hong Kong.

Obstetric outcome
All subjects were followed until delivery. The gestational age at delivery, mode of delivery, obstetric complications if any, and neonatal outcomes were recorded. The study subjects and clinical staff responsible for their care were blinded to the CRH results. For those subjects who subsequently delivered in another hospital, the obstetric information was obtained by telephone from the subject or via contact with staff in the other hospitals.

Pre-eclampsia was defined on the basis of a diastolic blood pressure recording greater than 110 mm Hg on one occasion or greater than 90 mm Hg on two or more occasions at least 4 h apart with the presence of significant proteinuria in subjects with no history of hypertension. Significant proteinuria was defined as proteinuria greater than 0.3 g/day or >=2+ on dipstix testing in two clean-catch mid-stream urine specimens collected at least 4 h apart.

Statistics
Statistical analyses were performed using the Statistical Package for the Social Sciences Version 9.0 (SPSS Inc., IL, USA). As both CRH and AFP concentrations vary with gestational age, their results were corrected for gestational age by expression as multiples of the median (MoM) of the normal pregnancies at that gestational age to allow legitimate comparison between normal and PE groups. Median concentrations of AFP at different gestational ages for the local population were used for the adjustment of AFP concentrations. The median concentrations of the CRH concentrations at different gestational ages were calculated from this cohort of subjects since such values for the local population were not otherwise available. The difference in the CRH and AFP concentrations in subjects who had PE and normal uncomplicated term pregnancies were examined using the Mann–Whitney U test. A value of P < 0.05 was considered statistically significant. The ability of mid-trimester CRH or AFP concentrations to predict PE was assessed by receiver operating characteristic (ROC) curves. The sensitivity, specificity, positive predictive value, negative predictive value, and the positive likelihood ratio for a threshold value of >=2.0 MoM were then calculated and discussed.

The test performance of combining the mid-trimester CRH and AFP concentrations was also assessed using ROC curves. Using logistic regression, a probability P for developing PE was formulated based on CRH and AFP concentrations {P = 1/[1 + exp(5.8167–0.3981xCRH–1.1019xAFP)]}. The P value for each individual subject was calculated and a ROC curve was then plotted according to the P value in predicting PE. The optimal sensitivity, specificity, positive predictive value, negative predictive value, and the positive likelihood ratio were then reported.


    Results
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Reference
 
A total of 502 subjects had blood sampled for CRH between 24 to 42 weeks of gestation, 13 (2.6%) of whom had PE diagnosed prior to the time of sampling. This incidence of PE was comparable to that in the local population, 2.9% (Hong Kong territory-wide audit, 1996). Twenty-six subjects who had an uncomplicated antenatal course and delivered at term were selected at random to match against the PE group for age and gestation at the time of blood sampling. A summary of the characteristics of the normal and PE groups is shown in Table IGo. As expected, the PE group had significantly shorter gestational length and lower birth weights. Seven of 13 subjects underwent iatrogenic preterm delivery before 37 weeks of gestation because of PE while one subject had spontaneous preterm delivery at 36.1 weeks. The CRH concentrations for all 39 subjects were detectable and ranged from 43 to 10 043 pmol/l. As the concentrations increased exponentially as gestation advanced, each CRH result was corrected for gestational age by expression as multiples of the median (MoM) of the normal pregnancies at that gestational age to allow legitimate comparison between normal and PE groups. After correction, the median CRH concentrations for the PE and the normal groups were 6.53 MoM [inter-quartile range (IQR): 4.05–11.36] and 1.00 MoM (IQR: 0.67–1.26) respectively (Figure 1Go). There was a statistical difference in the CRH concentrations between the two groups (P < 0.001, Mann–Whitney U test).


View this table:
[in this window]
[in a new window]
 
Table I. A summary of the characteristics of the normal and pre-eclamptic (PE) groups in the third trimester of pregnancy
 


View larger version (12K):
[in this window]
[in a new window]
 
Figure 1. Third trimester maternal corticotrophin-releasing hormone (CRH) concentrations for normal and pre-eclamptic (PE) groups. MoM = multiples of the median.

 
A total of another 1021 subjects were sampled for CRH and AFP in the mid-trimester between 15–22 weeks. All were free of pregnancy complications at the time of blood sampling. The mean age of this cohort was 30.8 years (SD 4.9) and 475 subjects (46.8%) were nulliparous. The mean gestational age at blood sampling was 18.1 weeks (SD 1.3; range: 15.0–22.3). A total of 748 subjects (73.3%) had an uncomplicated antenatal course and delivered at term (normal group) while another 21 subjects (2.1%) developed PE in their pregnancies. Six subjects (28.6%) had iatrogenic preterm delivery because of PE. None of the study subjects developed eclampsia.

A summary of the characteristics of the normal and PE groups is shown in Table IIGo. There was no difference in the maternal age between the two groups but a significantly larger proportion of subjects in the PE group were nulliparous. The mean gestational age at delivery was significantly lower for the PE group due to a higher incidence of preterm delivery because of the disease. Babies born to the PE group were significantly lighter and a significantly larger proportion were `small-for-dates', which was defined as birth weight below the 10th centile for the respective gestational age at delivery.


View this table:
[in this window]
[in a new window]
 
Table II. A summary of the characteristics of the normal and PE groups in the second trimester of pregnancy
 
Forty-nine subjects (4.8%) had concentrations of CRH below the limit of detectability, i.e. 1.35 pmol/l, and the CRH concentrations for the other subjects ranged from 1.35 to 119.55 pmol/l. All subjects had detectable concentrations of AFP, which ranged from 4.19 to 78.63 ng/ml. After adjustment of the concentrations according to the gestational age at sampling, the median CRH concentrations for the normal and PE groups were 1.00 (IQR: 0.55–1.78) MoM and 1.77 (IQR: 1.29–3.97) MoM respectively (Figure 2aGo). There was a statistical difference in the CRH concentrations between the two groups (P = 0.001, Mann–Whitney U test). The median AFP concentrations for the normal and PE groups were 0.91 (IQR: 0.73–1.18) MoM and 1.17 (IQR: 0.93–1.35) MoM respectively (Figure 2bGo). There was also a significant difference in the AFP concentrations between the two groups (P = 0.017, Mann–Whitney U test). The median CRH concentration for the six PE subjects who required preterm delivery was 2.49 MoM (IQR: 1.33–6.09) and was higher than that of the 15 PE subjects who delivered at term (median CRH = 1.56 MoM; IQR: 1.43–3.72). This difference did not reach statistical significance. However, a significant negative correlation was demonstrated between gestation at delivery and the mid-trimester CRH concentrations among the PE subjects (r = –0.54, P = 0.01). The median AFP concentrations for the preterm PE and term PE subjects were 1.28 MoM (IQR: 1.14–2.12) and 1.04 MoM (IQR: 0.86–1.32) respectively. This difference did not reach statistical significance. There was no significant correlation between AFP concentrations and gestation at delivery among PE subjects. Furthermore, there was no correlation between mid-trimester CRH or AFP concentrations and the severity of PE.



View larger version (12K):
[in this window]
[in a new window]
 
Figure 2. (a) Mid-trimester maternal CRH concentrations and (b) alpha-fetoprotein (AFP) concentrations in normal and pre-eclamptic (PE) groups.

 
Figure 3Go shows the ROC curves for using either CRH or AFP alone or a combination of both hormones as a predictor for PE. The areas under the curves for CRH or AFP alone or the combination of both were 0.70 (SD 0.06), 0.65 (SD 0.06) and 0.72 (SD 0.06) respectively. All three were significantly different from the area of no accuracy (0.5). If a threshold of 2.0 MoM was used, the sensitivity, specificity, positive predictive value, and negative predictive value for the mid-trimester CRH test in prediction of PE were 38.1, 79.9, 3.8 and 98.4% respectively. The positive likelihood ratio was 1.9 [95% confidence interval (CI): 1.1–3.3]. Regarding AFP, the sensitivity, specificity, positive predictive value, and negative predictive value for PE were 4.8, 98.6, 6.7 and 98% respectively for the same threshold value. The positive likelihood ratio was 3.4 (95% CI: 0.5–24.7). This showed that mid-trimester AFP is a very weak predictor for PE.



View larger version (14K):
[in this window]
[in a new window]
 
Figure 3. Receiver operating characteristic curves for prediction of pre-eclampsia.

 
After combining both CRH and AFP values, the best performance according to the maximum Youden index produced sensitivity, specificity, positive predictive value and negative predictive value of 61.9, 76.2, 5.2 and 99.0% respectively. The positive likelihood ratio was 2.6 (95% CI: 1.8–3.7). To allow comparison with the performance using CRH alone, at the same specificity of 79.9%, the sensitivity, positive predictive value and negative predictive value were 52.4, 5.2 and 98.8% respectively. The likelihood ratio remained 2.6 (95% CI: 1.7–4.0). Thus, the combination of mid-trimester CRH and AFP values improved the sensitivity, positive and negative predictive values, and the positive likelihood ratio for identification of PE subjects when compared with using the CRH test alone.


    Discussion
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Reference
 
The first part of this study has confirmed that maternal plasma CRH concentrations were significantly elevated in Chinese pregnancies complicated by pre-eclampsia. This is in agreement with that previously reported in a predominantly Caucasian population (Wolfe et al., 1988Go; Laatikainen et al., 1991Go; Emanuel et al., 1994Go; Perkins et al., 1995Go). The second part of the study has demonstrated that the mid-trimester maternal CRH concentrations were also significantly elevated in subjects who were destined to develop pre-eclampsia subsequently. This is in agreement with the concept that the underlying pathological changes in PE could have already occurred in this early period of pregnancy.

The reason for elevation of maternal CRH concentrations in subjects with pre-eclampsia is unknown. In human pregnancy, CRH is of placental origin and is secreted into both maternal and fetal circulations (Majzoub et al., 1999Go). CRH has been shown to exhibit a vasodilatory effect on the human feto-placental circulation and therefore may be a regulator of placental vascular tone (Clifton et al., 1994Go). Peripheral administration of CRH to human subjects has also been shown to cause vasodilatation and drop of blood pressure (Hermus et al., 1987Go). In the presence of hypoperfusion of the fetoplacental unit as in the case of PE, the elevation of CRH concentrations could be a protective response to placental ischaemia. Alternatively, the placental CRH may act on the fetal hypothalamic– pituitary–adrenal axis to cause increased production of cortisol, which is important for fetal maturation. A positive feedback loop between the fetal hypothalamic–pituitary–adrenal axis and placental CRH has been suggested by Robinson et al. (1988). It has been shown that the umbilical-cord plasma cortisol and CRH concentrations were both significantly elevated in neonates born to mothers with pre-eclampsia (Goland et al., 1995Go). It has also been reported that pulmonary maturation was more advanced in neonates born to women with pregnancy induced hypertension (Bent et al., 1982Go). The current data have shown that subjects with higher mid-trimester CRH concentrations in the PE group tended to require earlier delivery. This was mostly related to earlier onset of disease. It is possible that placental CRH production is elevated early in pregnancies with pre-eclampsia as a protective mechanism better to prepare the fetus for extrauterine life. However, we cound not demonstrate any correlation between maternal CRH concentrations and the severity of hypertension.

Of greater clinical relevance is the utility of the measurement for identification of those at high risk for PE. Although the data in this study showed that CRH measured in the mid-trimester could be used for prediction of PE, the performance was far from satisfactory. When a threshold of 2.0 MoM was used, the sensitivity, specificity, positive predictive value, negative predictive value and positive likelihood ratio were 38.1, 79.9, 3.8, 98.4%, and 1.9 respectively when the background prevalence was 2.1%. The performance of the CRH test in the mid-trimester was worse than when it was used for prediction of spontaneous preterm delivery (Leung et al., 1999Go).

Maternal plasma AFP concentrations were analysed to see if a combination of AFP and CRH concentrations would improve detection of high-risk subjects. The current data confirmed that mid-trimester maternal AFP concentrations were significantly elevated in subjects who would develop PE in the subsequent course of pregnancies. However, when used alone for prediction, mid-trimester maternal AFP concentrations were not clinically useful. Most of the studies on AFP and PE were case-controlled studies (Moore and Redman, 1983Go; Walters et al., 1985Go; Williams et al., 1992Go; Yaron et al., 1999Go) and the association of high mid-trimester maternal AFP concentrations with a higher risk of developing PE was well established. However, the clinical utility of the test for the purpose of prediction of PE can only be reflected in a prospective study. Milunsky et al. (1989) reported the AFP concentrations in 13 486 singleton pregnancies and showed that although high AFP was associated with a 2.3-fold increase of developing PE, the performance of the test for predicting PE was poor. The sensitivity, specificity, positive predictive value and negative predictive value were 8.9, 96.0, 3.6 and 98.4% respectively when the background prevalence was 1.7% (Milunsky et al., 1989Go). The results of the current study were in agreement with theirs in that the sensitivity and positive predictive value were very low.

The combination of AFP and CRH, however, offered increased detection of at-risk subjects. At the same specificity of 79.9% as when CRH >=2.0 MoM was used, the sensitivity was increased to 52.4% from 38.1%. The likelihood ratio was also increased to 2.6 from 1.9. However, this small increase in the likelihood ratio does not suggest the combination would be of great clinical value.

There is recent scientific evidence to suggest that poor placentation is not the only underlying pathology leading to pre-eclampsia (Ness and Roberts, 1996Go). Rather, the clinical syndrome is suggested to represent the end result of heterogeneous causes which lead to endothelial dysfunction or exaggerated maternal inflammatory response (Ness and Roberts, 1996Go; Redman et al., 1999Go). Although this remains hypothesis, it could explain why there is so far no early predictor that can accurately detect pre-eclampsia. The current findings are of no exception in this regard.

In summary, significantly elevated maternal CRH concentrations were demonstrated in subjects suffering from PE in the third trimester of pregnancy. This elevation of maternal CRH concentrations could be detected as early as in the mid-trimester before clinical signs of the disease have appeared. However, when used for prediction of pre-eclampsia, neither the CRH nor AFP concentrations alone in the mid-trimester had strong predictive value. Although a combination of both tests improved the predictive performance, the small improvement did not render the testing clinically useful.


    Acknowledgments
 
We thank Miss Diane Hazell and Mr Raydi Hui for laboratory work; the midwifery colleagues of the Department of Obstetrics and Gynaecology, Prince of Wales Hospital, Hong Kong for their assistance in collection of data; and Dr T.K.Lau, for helpful discussion. This project was supported by University Grant Committee of The Chinese University of Hong Kong.


    Notes
 
6 To whom correspondence should be addressed at: Department of Obstetrics and Gynaecology, The Chinese University of Hong Kong, The Prince of Wales Hospital, Hong Kong, China SAR. E-mail: dannytnleung{at}cuhk.edu.hk Back


    Reference
 Top
 Abstract
 Introduction
 Materials and methods
 Results
 Discussion
 Reference
 
Bent, A., Gray, J., Luther, E. et al. (1982) Assessment of fetal lung maturity: relationship of gestational age and pregnancy complications to phosphatidyl glycerol concentrations. Am. J. Obstet. Gynecol., 142, 664–669.[ISI][Medline]

Clifton, V.L., Read, M.A., Leitch, I.M. et al. (1994) Corticotrophin-releasing hormone-induced vasodilation in the human fetal placental circulation. J. Clin. Endocrinol. Metab., 79, 666–669.[Abstract]

Emanuel, R.L., Robinson, B.G., Seely, E.W. et al. (1994) Corticotrophin releasing hormone concentrations in human plasma and amniotic fluid during gestation. Clin. Endocrinol., 40, 257–262.[ISI][Medline]

Goland, R.S., Tropper, P.J., Warren, W.B. et al. (1995) Concentrations of corticotrophin-releasing hormone in the umbilical-cord blood of pregnancies complicated by pre-eclampsia. Reprod. Fertil. Dev., 7, 1227–1230.[ISI][Medline]

Hermus, A.R., Pieters, G.F., Willemsen, J.J. et al. (1987) Hypertensive effects of ovine and human CRFs in man. Eur. J. Clin. Pharmacol., 31, 531–534.[ISI][Medline]

Hong Kong College of Obstetricians and Gynaecologists (1996) Territory-wide Audit in Obstetrics and Gynaecology. Hong Kong College of Obstetricians and Gynaecologists, Hong Kong.

Laatikainen, T.J., Virtanen, T., Kaaja, R. et al. (1991) Corticotrophin-releasing hormone in maternal and cord plasma in pre-eclampsia. Eur. J. Obstet. Gynecol. Reprod. Biol., 39, 19–24.[ISI][Medline]

Leung, T.N., Chung, T.K.H., Madsen, G. et al. (1999) Elevated mid-trimester maternal corticotrophin-releasing hormone concentrations in pregnancies destined to deliver preterm. Br. J. Obstet. Gynaecol., 106, 1041–1046.[ISI][Medline]

Liapi, C.A., Tsakalia, D.E., Panitsa-Faflia, C.C. et al. (1996) Corticotrophin-releasing hormone concentrations in pregnancy induced hypertension. Eur. J. Obstet. Gynecol., 68, 109–114.[ISI][Medline]

Lyail, F. and Greer, J.A. (1994) Pre-eclampsia: a multifaceted vascular disorder of pregnancy. J. Hypertens., 12, 1339–1345.[ISI][Medline]

Majzoub, J.A., McGregor, J.A., Lockwood, C.J. et al. (1999). A central theory of preterm and term labor: putative role for corticotropin-releasing hormone. Am. J. Obstet. Gynecol., 180, S232–241.[ISI][Medline]

Milunsky, A., Jick, S.S., Bruell, C.L. et al. (1989) Predictive values, relative risks, and overall benefits of high and low maternal serum alpha-fetoprotein screening in singleton pregnancies: New epidemiologic data. Am. J. Obstet. Gynecol., 161, 292–297.

Moore, M.P. and Redman, C.W.G. (1983) Case-control study of severe pre-eclampsia of early onset. Br. Med. J., 287, 580–583.[ISI][Medline]

Ness, R.B. and Roberts, J.M. (1996) Heterogeneous causes constituting the single syndrome of pre-eclampsia: a hypothesis and its implications. Am. J. Obstet. Gynecol., 175, 1365–1370.[ISI][Medline]

Perkins, A.V., Linton, E.A., Eben, F. et al. (1995) Corticotrophin-releasing hormone and corticotropin-releasing hormone binding protein in normal and pre-eclamptic human pregnancies. Br. J. Obstet. Gynaecol., 102,118–122.[ISI][Medline]

Petraglia, F., Florio, P., Benedetto, C. et al. (1996) High concentrations of corticotrophin-releasing factor (CRF) are inversely correlated with low concentrations of maternal CRF-binding protein in pregnant women with pregnancy induced hypertension. J. Clin. Endocrinol. Metab., 81, 852–856.[Abstract]

Redman, C.W.G. (1991) Current topic: pre-eclampsia and the placenta. Placenta, 12, 301–308.[ISI][Medline]

Redman, C.W.G., Sacks, G.P., Sargent, I.L. (1999) Pre-eclampsia: an excessive maternal inflammatory response to pregnancy. Am. J. Obstet. Gynecol., 180, 499–506.[ISI][Medline]

Roberts, J.M. and Redman, C.W.G. (1993) Pre-eclampsia: more than pregnancy-induced hypertension. Lancet, 341, 1447–1451.[ISI][Medline]

Robertson, W.B., Brosens, L. and Dixon, H.G. (1967) The pathological response of the vessels of the placental bed to hypertensive pregnancy. J. Pathol. Bacteriol., 93, 581–592.[ISI][Medline]

Robinson, B.G., Emanuel, R.L., Frim, D.M. et al. (1988) Glucocorticoid stimulates expression of corticotrophin-releasing hormone gene in human placenta. Proc. Natl. Acad. Sci. USA, 85, 5244–5248.[Abstract]

Walters, B.N.J., Lao, T., Smith, V. et al. (1985) Alpha-fetoprotein elevation and proteinuric pre-eclampsia. Br. J. Obstet. Gynaecol., 92, 341–344.[ISI][Medline]

Williams, M.A., Hickok, D.E., Zingheim, R.W. et al. (1992) Elevated maternal serum {alpha}-fetoprotein concentrations and midtrimester placental abnormalities in relation to subsequent adverse pregnancy outcomes. Am. J. Obstet. Gynecol., 167, 1032–1037.[ISI][Medline]

Wolfe, C.D.A., Patel, S.P., Linton, E.A. et al. (1988) Plasma corticotrophin-releasing factor in abnormal pregnancy. Br. J. Obstet. Gynaecol., 95, 1003–1006.[ISI][Medline]

Yaron, Y., Cherry, M., Kramer, R.L. et al. (1999) Second trimester maternal serum marker screening: maternal serum {alpha}-fetoprotein, ß-human chorionic gonadotrophin, estriol, and their various combinations as predictors of pregnancy outcome. Am. J. Obstet. Gynecol., 181, 968–974.[ISI][Medline]

Submitted on November 30, 1999; accepted on May 4, 2000.