Early Prediction of Preeclampsia in High-Risk Women

Olfat Gamil Shaker, M.D.,1 and Hany Shehata, M.D.2

Abstract

Background: The purpose of this study was to determine the role of the combined use of uterine artery Dopplervelocimetry (UADV) and maternal serum placental growth factor (PlGF), vascular endothelial growth factorreceptor-1 soluble fms-like tyrosine kinase-1 also called soluble (sVEGFR-1), and nitric oxide (NO) productsconcentrations for the prediction of preeclampsia in high-risk women and to compare these parameters betweenpatients with mild and severe preeclampsia.Methods: Patients at risk of preeclampsia (n = 112) were subclassified as having either severe (n = 38), mild(n = 17), or no preeclampsia (n= 57). Blood samples were obtained between 22 and 26 weeks of gestation.Doppler ultrasound of the uterine arteries was done at the time of blood sampling. Maternal serum PlGF andsVEGFR-1 concentrations were determined with enzyme-linked immunosorbent assay (ELISA). Nitric oxidecolorimetric assay was used also to measure NO products in the maternal blood.Results: Among patients with abnormal UADV, maternal serum sVEGFR-1, PlGF, and NO product concen-trations contributed significantly in the identification of patients destined to develop mild and severe pre-eclampsia. sVEGFR-1 (pg/mL) concentration followed by NO product concentration (lmol/L) were found to bethe best predictors for preeclampsia, with high sensitivity and specificity, followed by PlGF (pg/mL).Conclusions: Abnormal UADV and high concentrations of sVEGFR1 combined with low concentrations of PlGFand NO products may be used to predict the development of preeclampsia.

Introduction

Preeclampsia, a pregnancy-specific syndrome charac-terized by new-onset hypertension and proteinuria, is aconsiderable obstetrical problem and a significant source ofmaternal and neonatal morbidity and mortality.1 It has beenrecognized that women who endure preeclampsia are at agreater risk than nonpreeclamptic women for cardiovasculardisease (CVD).2 Although the pathophysiology of pre-eclampsia remains undefined, placental ischemia/hypoxia iswidely regarded as a key factor.3 Inadequate trophoblast in-vasion leading to incomplete remodeling of the uterine spiralarteries is considered to be a primary cause of placental is-chemia.4 The poorly perfused and hypoxic placenta is thoughtto synthesize and release increased amounts of such vasoac-tive factors as soluble fms-like tyrosine kinase 1 (sFlt-1) (alsocalled soluble vascular endothelial growth factor receptor-(sVEGFR-1)), cytokines, and possibly the angiotensin II (ANGII) type 1 receptor autoantibodies (AT1-AA).

46

Several lines of evidence support the hypothesis that theischemic placenta contributes to endothelial cell dysfunctionin the maternal vasculature by inducing an alteration in the

balance of circulating levels of angiogenic/antiangiogenicfactors, such as vascular endothelial growth factor (VEGF),placental growth factor (PlGF), and sVEGFR-1.710 Data sug-gest that circulating sVEGFR-1concentrations may presagethe clinical onset of preeclamptic symptoms.9,11,12 VEGF isprimarily recognized for its potent angiogenic and mitogeniceffects on endothelial cells. It exerts its actions mainly by tworeceptors, VEGFR-1 and VEGFR-2 (also known as Flt-1) andthe kinase domain region (Flk/KDR), respectively.13 A solu-ble and endogenously secreted form of VEGFR-1 is producedmainly in the placenta by alternative splicing and contains theextracellular ligand-binding domain but not the transmem-brane and cytoplasmic portions.14

sVEGFR-1, a circulating antiangiogenic protein that se-questers the proangiogenic proteins PlGF and VEGF, is in-creased before the onset of clinical disease in the circulation ofwomen with preeclampsia.15 sVEGFR-1 disrupts VEGF sig-naling either by binding VEGF and PlGF or by forming het-erodimers with the KDR receptor.16 Although sVEGFR-1 isnot a vasoconstrictor, it does significantly inhibit the dilatoryactions of both VEGF and PlGF in vitro, and chronic eleva-tions in circulating concentrations cause increased blood

Departments of 1Medical Biochemistry and Molecular Biology and 2Obstetrics & Gynecology, Faculty of Medicine, Cairo University, Cairo,Egypt.

JOURNAL OF WOMENS HEALTHVolume 20, Number 4, 2011 Mary Ann Liebert, Inc.DOI: 10.1089/jwh.2010.2378

539

pressure.5,17 Preeclampsia is strongly linked to an imbalancebetween proangiogenic (VEGF and PlGF) and antiangiogenic(sVEGFR-1) factors in the maternal circulation.5,9,10,18 It hasbeen reported that increased sVEGFR-1may have a predictivevalue in diagnosing preeclampsia because concentrationsseem to increase beforemanifestation of overt symptoms (e.g.,hypertension and proteinuria).9,15

Substantial evidence indicates that nitric oxide (NO) pro-duction is elevated in normal pregnancy and that this increaseappears to play an important role in the renal vasodilatationof pregnancy.19 NO synthase (NOS) inhibition in pregnantrats produces hypertension associated with renal vasocon-striction, proteinuria, intrauterine growth restriction, and in-creased fetal morbidity.20,21

Abnormal uterine artery doppler velocimetry (UADV) be-tween 22 and 25weeks of gestation has been called best test foridentification of patients destined to develop preeclampsia,compared with biochemical indicators in the maternal serum.Abnormal UADV results and abnormal maternal serum con-centrations of proangiogenic and antiangiogenic factors are riskfactors for the subsequent development of preeclampsia.22,23

Currently, there is no widely accepted screening test forprediction of preeclampsia in individual women. The devel-opment of an accurate biomarker for preeclampsia in high-risk women has the potential to substantially improve care byallowing closer prenatal monitoring, recognition of pre-eclampsia earlier in the disease course, expeditious adminis-tration of steroids for fetal lung maturity, and appropriateantihypertensive therapy. The aim of this study was to de-termine the utility of maternal serum concentrations of theangiogenic factor PlGF and the antiangiogenic factorsVEGFR-1, in combination with UADV, for prediction ofpreeclampsia in the midtrimester of pregnancy.

Subjects and Methods

A prospective cohort study was conducted between April2008 and January 2010. Patients were recruited from thoseattending the gynecology clinic at Kasr El Aini hospital, CairoUniversity. Inclusion criteria were pregnancy of < 24 weeksgestation at enrollment and at least one of the following riskfactors for preeclampsia: pregestational diabetes mellitus,maternal age 18 years, systemic lupus erythematosus (SLE),or prior history of preeclampsia.

Preeclampsia was diagnosed and subdivided into eithersevere or mild according to published guidelines.24 UADVwas performed at the time of blood sampling. The presence ofan early diastolic notch in the uterine arteries was determinedaccording to the criteria proposed by Bower et al.25 An ab-normal UADVwas defined as the presence of bilateral uterineartery notches or a mean pulsatility index of > 95th percentilefor the gestational age or both. Themean pulsatility indexwascalculated by measuring the pulsatility index of the right andleft uterine arteries. Primary pregnancy outcome was the di-agnosis of early onset preeclampsia or severe preeclampsia,and secondary outcomes included small for gestational age(SGA), premature delivery (PTD), and intrauterine growthretardation (IUGR).

Human PlGF and sVEGFR-1 assays

The concentrations of sVEGFR-1 were measured in all serawith an enzyme-linked immunosorbent assay (ELISA) (R&D

Systems). The inter and intraassay coefficients of variation(CV) were 4.8% and 6.9%, respectively. The sensitivity of theassay was 3.5 pg/mL. A specific and sensitive ELISA wasused to determine concentrations of PlGF in maternal serum(R&D Systems). The calculated interassay and intraassay CVswere 4.6% and 2.27%, respectively. The minimum detectabledose of PlGF is < 7pg/mL.

Nitrite assay

The nitric oxide colorimetric assay kit provides a conve-nient measure of total nitrate/nitrite in a simple two-stepprocess. The first step converts nitrate to nitrite using nitratereductase. The second step uses Griess reagents to convertnitrite to a deep purple azo compound. The amount of the azochromophore accurately reflects the NO amount in samples.The detection limit of the assay is approximately 0.1 nmolenitrite/well, or 1 lM (Abcam).

Statistical analysis

The data were coded and entered using the statisticalpackage SPSS version 12. The data were summarized usingdescriptive statistics: mean, standard deviation (SD), minimaland maximum values for continuous variables, and numberand percentage for categorical values. Statistical differencesbetween groups were tested using the chi-square test forcategorical variables, analysis of variance (ANOVA) for con-tinuous normally distributed variables, and the nonpara-metric Mann-Whitney test and Kruskal-Wallis test for notnormally distributed continuous variables. Correlations weredone to test for linear relations between continuous variables.Logistic regression analysis was done to test for significantpredictors for preeclampsia. Receiver operator characteris-tic (ROC) curves were constructed to evaluate the predic-tive potential of each biomarker for preeclampsia occurrenceand severity. p Values 0.05 were considered statisticallysignificant.

Results

One hundred twelve pregnant women were enrolled in thestudy; 57 did not develop preeclampsia, and 55 developedpreeclampsia (17 mild preeclampsia and 38 severe pre-eclampsia). All cases of preeclampsia had abnormal UADV.

Baseline characteristics and pregnancy outcomes areshown in Table 1. Subjects with mild and severe preeclampsiahad a higher body mass index (BMI) ( p1, p2< 0.001) and in-crease in gestational systolic blood pressure (SBP) and dia-stolic blood pressure (DBP) ( p1, p2< 0.001) compared withthose who did not develop preeclampsia (normotensivegroup). Also, gestational SBP and DBP were significantlyhigher in the severe group than in the mild group ( p3< 0.001).Patients with severe preeclampsia delivered at an earliergestational age, had a higher mean arterial pressure, and de-livered smaller infants than subjects in the control and mildgroups. A significant difference in all other pregnancy out-come parameters (birth weight, placental weight, gestationalage at delivery, mean pulsatile index, SGA, and IUGR) wasfound between the studied groups.

Table 2 shows a significant increase in mean sVEGFR-1levels in subjects who developed mild and severe pre-eclampsia compared with those who did not develop pre-

540 SHAKER AND SHEHATA

eclampsia ( p1, p2< 0.001), whereas no significant differencewas found between the mild and severe groups ( p3 > 0.05).The mean serum PlGF levels were significantly lower in thesevere group compared with those without preeclampsia andthe mild group ( p2, p3< 0.001). The mean levels of serum NOproducts were significantly decreased in both the mild andsevere groups compared with the normotensive group ( p1,p2 < 0.001), but no significant difference was found betweenthe mild and severe groups ( p3> 0.05).

We determined the ROC curve to examine the diagnosticperformance of maternal serum PlGF, sVEGFR-1, and NOproduct concentrations in identifying patients destined todevelop mild or severe preeclampsia. sVEGFR-1 (pg/mL)followed by NO product (lmol/L) concentration were foundto be best for diagnosis of preeclampsia, with high sensitivityand specificity, followed by PlGF (pg/mL) concentrations(Tables 3 and 4). A significant correlation was observed be-

tween concentrations of sVEGFR-1, PlGF, and NO productsand pregnancy outcome characteristics (Table 5).

Discussion

Considerable clinical evidence has accumulated that pre-eclampsia is strongly linked to an imbalance betweenproangiogenic (VEGF and PlGF) and antiangiogenic(sVEGFR-1) factors in the maternal circulation.9 Recent stud-ies have reported that increased sVEGFR-1 may have a pre-dictive value in diagnosing preeclampsia, as concentrationsseem to increase beforemanifestation of overt symptoms (e.g.,hypertension and proteinuria).26

The present study showed a significant increase in meanmaternal sVEGFR-1 levels in subjects who developed mildand severe preeclampsia compared with those who did notdevelop preeclampsia. The mean serum levels of maternal

Table 1. Characteristics and Pregnancy Outcomes of All Subjects

CharacteristicNormotensivesubjects (n = 57)

Mild preeclampsia(n = 17)

Severe preeclampsia(n = 38) p1a p2 p3

Maternal age 27.8 9.6b 29.4 7.3 27.3 8.4 > 0.05 > 0.05 > 0.05Parity 1.77 0.88 1.76 0.83 1.73 0.89 > 0.05 > 0.05 > 0.05BMI (kg/m2) 25.0 5 31 4 29.0 4.0 < 0.001* < 0.001* > 0.05Gestational SBP (mm Hg) 104 13 143 6 170 16 < 0.001* < 0.001* < 0.001*Gestational DBP(mm Hg) 66 9 94 4 109 13 < 0.001* < 0.001* < 0.001*Birth weight (kg) 3.54 0.24 3.08 0.29 2.20 0.25 < 0.001* < 0.001* < 0.001*Placental weight (g) 567.98 54.63 486.47 38.88 422.37 45.83 < 0.001* < 0.001* < 0.001*Gestational age atdelivery (weeks)

39.54 0.56 37.65 0.99 33.66 0.62 < 0.001* < 0.001* < 0.001*

Mean pulstile index 2.73 0.08 2.60 0.40 1.25 0.42 > 0.05 < 0.001* < 0.001*SGA 0 (0%)c 0 (0%) 6 (15.8%) > 0.05 < 0.001* < 0.001*IUGR 0 (0%) 1 (5.9%) 24 (63.2%) < 0.001* < 0.001* < 0.001*

ap1, between normotensive and mild preeclampsia groups; p2, between normotensive and severe preeclampsia groups; p3, between mildpreeclampsia and severe preeclampsia groups.

bmean standard deviation (SD).cnumber (%).*Significant.BMI, body mass index; DBP, diastolic blood pressure; IUGR, intrauterine growth retardation; SBP, systolic blood pressure; SGA, small for

gestational age.

Table 2. Measured Parameters Estimating Extent of Preeclampsia in All Studied Groups

Parameter/group Normotensive Mild preeclampsia Severe preeclampsia p valuea

PlGF (pg/mL) 407.55 149.82b 260.73 112.37 174.97 116.48 p1 > 0.05p2 < 0.001*p3 < 0.001*

sVEGFR-1 (pg/mL) 1101.70 157.86 4707.05 2134.32 5184.21 1652.60 p1 < 0.001*p2 < 0.001*p3 > 0.05

NO products (lmol/L) 117.17 28.42 43.25 23.84 38.70 12.53 p1 < 0.001*p2 < 0.001*p3 > 0.05

ap1, between normotensive and mild preeclampsia groups; p2, between normotensive and severe preeclampsia groups; p3, between mildpreeclampsia and severe preeclampsia groups.

bMean SD.*Significant.NO, nitric oxide; PlGF, placental growth factor; sVEGFR-1, soluble vascular endothelial growth factor receptor-1.

PREDICTION OF PREECLAMPSIA 541

PlGF were significantly decreased. Our sVEGFR-1 and PlGFfindings were consistent with results of the study by MooreSimas et al.27 They found that mean sVEGFR-1 levels weresignificantly higher in subjects who developed preeclampsiabefore 34 weeks compared with those without preeclampsia.At the same time, they found that themean PlGF levels tendedto be lower for subjects who developed preeclampsia com-pared with those without preeclampsia. They studied thesVEGFR-1/PlGF ratio as an index of antiangiogenic activitythat reflects changes in the balance between sVEGFR-1 andPlGF. They suggested that this ratio has been shown to bemore strongly associated with preeclampsia than eithermeasure alone in healthy women. They concluded that inhigh-risk women, serum sVEGFR-1 and the sVEGFR-1/PlGFratio are altered before preeclampsia onset and may be pre-dictive of preeclampsia.

In this study, ROC curves were constructed to describethe relationship between sensitivity and the false positive rateof serum PlGF, sVEGFR-1, and NO products in identify-ing patients destined to develop preeclampsia. sVEGFR-1(pg/mL) and NO products (lmol/L) were found to be thebest predictors for preeclampsia, with high sensitivity andspecificity, followed by PlGF (pg/mL), whereas in severepreeclampsia, sVEGFR-1 was the best predictor, followed byNO products, then PlGF. Logistic regression analysis indi-cated that maternal serum concentrations of sVEGFR-1 > 2005 pg/mL, NO products < 50.9 lmol/L and PLGF