early prediction of preeclampsia and iugr dissertation.pdf · early prediction of preeclampsia and...

Early Prediction of Preeclampsiaand IUGR

Koen L. Deurloo

The studies described in this thesis were performed at the Division of Prenatal Diagnosis and Therapy, Department of Obstetrics and Gynecology and the Department of ClinicalChemistry, VU University Medical Center, Amsterdam, the Netherlands.

Financial support for printing of this thesis was kindly provided by

Ferring PharmaceuticalsMedical DynamicsEsaote – Pie Medical Benelux B.V.Via MedicaMemidis Pharmabma bv (Mosos)J.E. Jurriaanse Stichting.

Cover and design: Willem Morelis, Jeffrey van der Toorn en Renée DeurlooTypesetting: Willem MorelisPrinted by: Wöhrmann Printserviceisbn 978 90 865 9466 5

© 2010, Koen L. Deurloo, Amsterdam, the Netherlands.No part of this thesis may be reproduced, stored or transmitted by any other form or by anymeans, without prior permission of the author.

vrije universiteit

Early Prediction of Preeclampsia and Intra Uterine Growth Restriction

academisch proefschrift

ter verkrijging van de graad Doctor aan

de Vrije Universiteit Amsterdam,

op gezag van de rector magnificus

prof.dr. L.M. Bouter,

in het openbaar te verdedigen

ten overstaan van de promotiecommissie

van de faculteit der Geneeskunde

op woensdag 23 juni 2010 om 13.45 uur

in de aula van de universiteit,

De Boelelaan 1105

door

Koen Laurens Deurloo

geboren te ’s-Gravenhage

Promotoren: prof.dr. J.M.G. van Vugt

prof.dr. M.A. Blankenstein

Copromotor: dr. A.C. Bolte

Voor mijn meisjes

Voor mijn ouders

Contents

chapter 1Introduction and outline

chapter 2Reproducibility of three-dimensional ultrasound measurementsof fetal and placental volume at gestational ages of 11-18 weeks

chapter 3Clinical use of three-dimensional ultrasound fetal and placentalvolumes

chapter 4Longitudinal Doppler measurements of spiral artery blood flowin relation to uterine artery blood flow

chapter 5Color Doppler ultrasound of spiral artery blood flow for prediction of hypertensive disorders and intra uterine growthrestriction: a longitudinal study

chapter 6adam12s and pp13 as first trimester screening markers for adverse pregnancy outcome

chapter 7Decreased plasma levels of metastin in early pregnancy are associated with small for gestational age neonates

chapter 8General discussion, conclusions and future prospects

chapter 9SummarySamenvatting

List of co-authorsList of publicationsDankwoord

7

19

33

41

53

67

81

93

108

113

118

119

120

Introduction and outline

chapter 1

Preeclampsia (pe) and intrauterine growth restriction (iugr) are causes of con-sid erable maternal and fetal morbidity and mortality with partly understood etiology.1,2 It is widely believed that during the first half of pregnancy, trophoblasticinvasion is responsible for remodelling of the spiral arteries into dilated, inelasticvessels capable of supplying large amounts of blood to the placenta and the devel -oping fetus. This physiological process is responsible for the development of afunctional uteroplacental circulation.3,4 Histological studies show that in uncom-plicated pregnancies the invasion of the trophoblastic cells extends to the decidualand myometrial segments, changing the spiral arteries to low-resistance vesselscapable of supplying high amounts of blood to the placenta.4-6

Impaired remodeling due to shallow trophoblastic invasion is associated withmaintenance of high resistance and low flow spiral arteries as comparable to thenon-pregnant uterus and related to possible subsequent development of pe withor without iugr and isolated iugr.4,7-9 Various authors have reported that whentrophoblastic invasion is impaired, the invasion is usually restricted to the decidualsegments leaving the myometrial segment of the spiral arteries unchanged andretaining their musculo-elastic wall.5

Early monitoring of trophoblastic invasion may increase our knowledge aboutplacental development. It enables us to identify women with high risk for devel-oping obstetric complications, such as pe and/or iugr secondary to uteroplacentaldysfunction, who require more intensive surveillance during pregnancy. In addi-tion, early identification of high risk women might help to discriminate betweeniugr fetuses from constitutional small-for-gestational-age (sga) fetuses.

Reported prevalence of pe in an unselected population is dependent on the diag-nostic criteria used and ranges from 0.5% for severe pe to 2-5% overall.10 Inci-dence of iugr is dependent of the definition of fetal weight used: less than 2.3, 5or 10th percentile or fetal weight two standard deviations below the mean forgestational age. In most literature, a fetal weight below the tenth percentile is theaccepted definition of iugr.11,12 Since pe and iugr are complications with relativelylow incidence in western population, a screening test should therefore accuratelyidentify high risk women for developing pe and iugr.13

Pre-pregnancy discrimination of women into low an high risk groups can beaccomplished by screening of various pre-pregnancy risk factors, such as nullipar-ity, pre-existing maternal diseases such as hypertension, diabetes mellitus, highbody mass index or maternal age or a history of obstetric complications, such aspreeclampsia or iugr.14-15 However, these pre-pregnancy risk factors have lowpredictive values and need to be supported by other screening tests, especiallyduring pregnancy.

EARLY PRED ICTION OF PREECLAMPS IA AND IUGR8

A good screening test for pe and isolated iugr should be safe, simple, quick andinexpensive and patient’s acceptability should be taken into account as well. Earlyidentification allows early intervention and treatment for high risk women, whichmay decrease maternal and fetal morbidity and improve outcome, although atthis moment secondary preventive options are very limited and currently no cureexcept delivery exists. High sensitivity is more preferable than high specificity be-cause minimizing false negatives over false positives is advantageous consideringthe benefits and harms due to unnecessary treatment or intervention. But a goodtest should preferably incorporate both high sensitivity as well as high specificity.

Over the last decades, numerous authors have studied the potential of first andsecond trimester maternal and fetal marker screening to identify impaired placen-tal development and subsequent pregnancy complications, such as pe and isolatediugr. Various potent predictive markers have been studied, such as Dopplermeasurements, three dimensional ultrasound volume measurements and maternalserum markers.16,17 However, large and even controversial differences in screeningperformance are reported.

Doppler Ultrasound

Doppler ultrasound is a non-invasive technique, which can easily be used forevaluation of maternal and fetal hemodynamics. It was first described by Fitz -gerald in 1977, who used a standard ultrasound device combined with a Dopplerultrasound beam.18 His recordings were analyzed by headphone and allowed himto observe the fetal umbilical arteries. Since then, Doppler ultrasound techniquehas rapidly developed into a quick and easy way to study fetal and maternal cir-culation.

Uterine arteryMost reports about the use of Doppler ultrasound for screening purposes dur-ing pregnancy focus on the uterine arteries (ua), which are the main arteries forblood supply of the pregnant uterus. They are believed to reflect the decrementin impedance to flow in spiral arteries during trophoblastic invasion.

In 1983, Campbell was the first to study ua (arcuate) Doppler measurements insecond and third trimester in healthy women with uncomplicated pregnancies.19

He describes a unique characteristic shape: continuous positive flow with highend-diastolic velocities, which makes the ua’s easy to identify. In addition, he reports

9CHAPTER 1 INTRODUCTION AND OUTLINE

that in pregnancies complicated by proteinuric hypertension and iugr the ua

vascular resistance was increased in half of them. Two years later these findingswere confirmed by Trudinger et al. in small-for-gestational-age infants in whomover half had serial abnormal flow indices from 20 till 40 weeks of gestation.20

Although generally assumed, it was not until 1993 that Olofsson in a case-controlstudy examined placental bed biopsies taken during caesarean section in compli-cated and uncomplicated pregnancies and he linked increased ua resistance toimpaired trophoblastic invasion.21 A year later, Meekins also found differences introphoblastic invasion of the myometrial spiral arteries in normal compared tosevere preeclamptic pregnancies while studying placental bed biopsies.6 In thelate nineties, Lin and Sagol confirmed this relationship of increased ua Dopplerindices and impaired trophoblastic invasion in pe and iugr complicated pregnan-cies.22,23

Soon after the discovery of ua Doppler us measurements, its use has beenevaluated as an early predictive marker for adverse pregnancy outcome by dif-ferent authors. Campbell was one of the first authors to publish results on theuse of ua Doppler screening for identification of high risk pregnancies, such ashypertensive disorders, fetal growth restriction and foetal asphyxia.19 Screeningperformance was moderate with a sensitivity of 68% and a specificity of 69%, butit seemed a promising tool for early screening for high risk pregnancies.

Over the years many reports on screening performance of ua artery Dopplermeasurements have been published with considerable different outcomes, rang-ing form very good to poor predictive values. Although diagnostic accuracy ishampered by different tests, different populations and different gestational agesat measurement, most reports show high sensitivity with moderate to poorspecificity and low predictive scores for hypertensive disorders and intrauterinegrowth restriction.24-34 However, few authors report good predictive values withsensitivity levels over 70%.35-37

In 2008 a large systematic review was published in which nearly 80.000 womenwere screened for pe and over 40.000 women were screened for iugr with theuse of ua Doppler measurements.38 In low risk women the overall risk for pe wasbest predicted by second trimester measurement: pulsatility index with presenceof bilateral notching: high specificity 99% and moderate sensitivity of 23%. In highrisk women, the overall risk for pe was best predicted by ua pulsatility index andbilateral notching in the second trimester: specificity 99%, sensitivity 19%. In-creased ua pulsatility index and bilateral notching was also the best predictor for


severe and overall iugr with specificity of 99% and 98% and sensitivity of 12%and 23% respectively.

Routine use of ua Doppler us screening seems considering the overall testcharacteristics with high false positive rates and the absence of adequate therapynot (yet) a useful option, however its high negative predictive values might identifya low risk population which can be excluded from increased surveillance.39

Spiral arteryAfter branching from the hypogastric artery, the uterine arteries further divideinto basal, radial and finally into spiral arteries. They were discovered by accidentas arteriae convolutae in the late 18th century by the Hunter brothers, but it wasuntil the twentieth century that spiral arteries were linked to trophoblastic inva-sion by Brosens in histological studies.5

Doppler us measurements of these small spiral arteries were technically notpossible until the early nineties when Color Doppler was introduced. In 1991

Pijnenborg et al. linked impaired sa modification to impaired trophoblastic invasionand Jurkovic et al. published a study concerning spiral artery hemodynamics in un-complicated pregnancies and describes a characteristic waveform and progres-sively decreasing indices of impedance to flow with advancing gestational age.9,40

Doppler ultrasound measurements in sa seemed to correspond with (patho)physiological change in sa in a histological study of early pregnancies.41

Flow velocity waveforms in normal and complicated pregnancies were stud-ied by several authors in matched-pairs and cross sectional studies. However, nolongitudinal data has been reported in literature.

Screening performance of sa in prediction of complicated pregnancies wasassessed in a few studies: Locci et al. in 1993 did not consider sa flow indices in pe

and iugr complicated pregnancies a good way of monitoring and found moderatepredictive values (sensitivity 53%, specificity 40%, positive predictive value 65%and negative predictive value 33%).42 Two other studies report increased secondand third trimester Doppler indices in pe and iugr complicated pregnancies; Murakoshi found very good predictive values for pih and iugr with sensitivity85% and specificity of 91%.43 Matijevic confirmed these findings in a case-controlstudy with preeclamptic pregnancies compared to uncomplicated pregnancies.44

First trimester sa Doppler indices are considered poor predictors of adversepregnancy outcome by Hung et al., with sensitivity and specificity close to 50%.45


Serum markers

For early detection of women at high risk of developing pe or iugr, a wide varietyof serum markers have been evaluated, including markers of renal dysfunction,endothelial dysfunction and oxidative stress, markers for aneuploidy as well asplacental peptide hormones.46,47 Two large systematic reviews of screening testsfor preeclampsia in 2002 and 2004 reported that there is still no clinically usefulscreening test to predict the development of preeclampsia, probably due to itsmultifactorial etiology and its low incidence in a general population.48,49

The search for predictive markers is still continuing. Since 2004 many newpromising markers have been evaluated, such as angiogenesis factors (PlGF, vegf,sFlt, sEng), placental proteins (adam12s, pp13 and ptx3), coagulation factors (p-selectin) and cell-free fetal dna. However, the reliability of these markers is in-consistent between different studies and although several markers might offerthe potential to be used as a predictive tool, better predictive values are to beexpected from combinatory analysis of multiple markers.

Promising multiple markers screening combine sonographic markers, serummarkers and/or Doppler measurements for prediction of pe, such as first tri -mester pp13, papp-a and second trimester ua pulsatility index by Spencer,49 sFltand PlGF with second trimester ua pulsatility index,50

papp-a, PlGF and ua pi51 or

first trimester ua pulsatility index and placenta volume.52 Multiple serum markerscreening such as soluble endoglin, PlGF and sFlt 53 or in combination with tgf-beta-154, or papp-a, beta-hCG and adam12

55 have been subject of evaluation as well.For the prediction of iugr, a combination of PlGF, vegf, vegfr-1 and placentalvolume,56 as well as first trimester ua pi and papp-a measurements57 have beenstudied. Predictive values differ considerably between different studies and al-though good screening performance has been reported, there is need for pro spec -tive, large, multicenter, multi-ethnic studies to confirm these promising results.

Three-dimensional ultrasound volume measurements

Three-dimensional ultrasound (3dus) is a promising tool in early prediction ofcomplicated pregnancies since its introduction in 1980. Due to the rapid devel-opment of computer processor technology, 3dus have been used for quick andaccurate assessment of various volume measurements, such as gestational sac58 andyolk sac59 or fetal organs, such as liver,60 lungs61 and kidneys.62 Although broadlyused in many aspects of prenatal screening and diagnosis, its role in (early) pre-


diction or identification of pregnancies complicated by pe or iugr remains modest.Several authors have examined the use of placental volume measurements withdifferent outcomes.63-68

Aim of Thesis

The aim of this thesis was to study the predictive value of different maternal serumand sonographic markers for early prediction of preeclampsia and isolated in-trauterine growth restriction in a high-risk population.

chapter 2 gives a detailed assessment of the reproducibility of three dimension-al ultrasound fetal and placental volume measurements, in which intra- and inter-observer agreement was studied. In chapter 3 the clinical relevance of threedimensional ultrasound measurements of placental and fetal volumes in predic-tion of preeclampsia and isolated intrauterine growth restriction are reviewed.

The assumed relationship between Doppler measurements of uterine andspiral arteries is discussed in chapter 4 and in chapter 5 the value of measuringspiral arteries in prediction of preeclampsia, pregnancy induced hypertensionand isolated intra uterine growth restriction was evaluated. Promising maternalserum markers Metastin, Placental Protein 13 and adam12s with predictive valuesfor preeclampsia, pregnancy induced hypertension and intrauterine growth re-striction were analyzed in chapter 6 and chapter 7.

This thesis ends with a general discussion and future prospects in chapter 8.A summary is given in chapter 9.


References

1 Sibai B, Dekker G, Kupferminc M. Pre- eclampsia. Lancet 2005; 9461: 785-799.

2 Khan KS, Wojdyla D, Say L, Gulmezoglu AM,Van Look PF. Who analysis of causes of maternal death: a systematic review. Lancet2006; 9516: 1066-1074.

3 Pijnenborg R, Bland JM, Robertson WB, Brosens I.Uteroplacental arterial changes related to interstitial trophoblast migration in early humanpregnancy. Placenta 1983; 4: 397-413.

4 Lyall F. Priming and remodelling of human placental bed spiral arteries during pregnancy – a review. Placenta 2005; S31-S36.

5 Brosens I. A study of the spiral arteries of thedecidua basalis in normotensive andhypertensive pregnancies. J Obstet Gynaecol BrCommonw 1964; 222-230.

6 Meekins JW, Pijnenborg R, Hanssens M, McFadyen IR, van Asshe A. A study of placentalbed spiral arteries and trophoblast invasion innormal and severe pre-eclamptic pregnancies.Br J Obstet Gynaecol 1994; 8: 669-674.

7 Brosens IA, Robertson WB, Dixon HG. Therole of the spiral arteries in the pathogenesis ofpreeclampsia. Obstet Gynecol Annu 1972; 177-191.

8 Sheppard BL and Bonnar J. An ultrastructuralstudy of utero-placental spiral arteries inhypertensive and normotensive pregnancy andfetal growth retardation. Br J Obstet Gynaecol1981; 7: 695-705.

9 Pijnenborg R, Anthony J, Davey DA, Rees A,Tiltman A, Vercruysse L, van AA. Placental bedspiral arteries in the hypertensive disorders ofpregnancy. Br J Obstet Gynaecol 1991; 7: 648-655.

10 Saftlas AF, Olson DR, Franks AL, Atrash HK,Pokras R. Epidemiology of preeclampsia andeclampsia in the United States, 1979-1986. Am J Obstet Gynecol 1990; 2: 460-465.

11 Battaglia FC and Lubchenco LO. A practicalclassification of newborn infants by weight andgestational age. J Pediatr 1967; 2: 159-163.

12 Seeds JW and Peng T. Impaired growth and riskof fetal death: is the tenth percentile theappropriate standard? Am J Obstet Gynecol 1998;4: 658-669.

13 Villar J, Say L, Shennan A, Lindheimer M, DuleyL, Conde-Agudelo A, Merialdi M.Methodological and technical issues related tothe diagnosis, screening, prevention, andtreatment of pre-eclampsia and eclampsia. Int J Gynaecol Obstet 2004; S28-S41.

14 Milne F, Redman C, Walker J, Baker P, Bradley J,Cooper C, de Swiet M, Fletcher G, Jokinen M,Murphy D, Nelson-Piercy C, Osgood V, RobsonS, Shennan A, Tuffnell A, Twaddle S, Waugh J.The pre-eclampsia community guideline(precog): how to screen for and detect onsetof pre-eclampsia in the community. bmj 2005;7491: 576-580.

15 Duckitt K and Harrington D. Risk factors forpre-eclampsia at antenatal booking: systematicreview of controlled studies. bmj 2005; 7491:565-

16 Poon LC, Maiz N, Valencia C, Plasencia W,Nicolaides KH. First-trimester maternal serumpregnancy-associated plasma protein-A andpre-eclampsia. Ultrasound Obstet Gynecol 2009;1: 23-33.

17 Spencer K, Cowans NJ, Nicolaides KH. Lowlevels of maternal serum papp-a in the firsttrimester and the risk of pre-eclampsia. PrenatDiagn 2008; 1: 7-10.

18 FitzGerald DE and Drumm JE. Non-invasivemeasurement of human fetal circulation usingultrasound: a new method. Br Med J 1977; 6100:1450-1451.

19 Campbell S, Diaz-Recasens J, Griffin DR,Cohen-Overbeek TE, Pearce JM, Willson K,Teague MJ. New doppler technique forassessing uteroplacental blood flow. Lancet1983; 8326 Pt 1: 675-677.

20 Trudinger BJ, Giles WB, Cook CM.Uteroplacental blood flow velocity-timewaveforms in normal and complicatedpregnancy. Br J Obstet Gynaecol 1985; 1: 39-45.

21 Olofsson P, Laurini RN, Marsal K. A high uterineartery pulsatility index reflects a defectivedevelopment of placental bed spiral arteries inpregnancies complicated by hypertension andfetal growth retardation. Eur J Obstet GynecolReprod Biol 1993; 3: 161-168.

22 Lin S, Shimizu I, Suehara N, Nakayama M, Aono T. Uterine artery Doppler velocimetry


in relation to trophoblast migration into themyometrium of the placental bed. ObstetGynecol 1995; 5 Pt 1: 760-765.

23 Sagol S, Ozkinay E, Oztekin K, Ozdemir N. The comparison of uterine artery Dopplervelocimetry with the histopathology of theplacental bed. Aust N Z J Obstet Gynaecol 1999;3: 324-329.

24 Steel SA, Pearce JM, Chamberlain G. Dopplerultrasound of the uteroplacental circulation as a screening test for severe pre-eclampsia withintra-uterine growth retardation. Eur J ObstetGynecol Reprod Biol 1988; 4: 279-287.

25 Steel SA, Pearce JM, McParland P, ChamberlainGV. Early Doppler ultrasound screening inprediction of hypertensive disorders ofpregnancy. Lancet 1990; 8705: 1548-1551.

26 Bower S, Schuchter K, Campbell S. Dopplerultrasound screening as part of routineantenatal scanning: prediction of pre-eclampsiaand intrauterine growth retardation. Br J ObstetGynaecol 1993; 11: 989-994.

27 Zimmermann P, Eirio V, Koskinen J, Kujansuu E,Ranta T. Doppler assessment of the uterine anduteroplacental circulation in the secondtrimester in pregnancies at high risk for pre-eclampsia and/or intrauterine growth retardation:comparison and correlation between differentDoppler parameters. Ultrasound Obstet Gynecol1997; 5: 330-338.

28 Irion O, Masse J, Forest JC, Moutquin JM.Prediction of pre-eclampsia, low birthweight for gestation and prematurity by uterine arteryblood flow velocity waveforms analysis in lowrisk nulliparous women. Br J Obstet Gynaecol1998; 4: 422-429.

29 Kurdi W, Campbell S, Aquilina J, England P,Harrington K. The role of color Dopplerimaging of the uterine arteries at 20 weeks’gestation in stratifying antenatal care. UltrasoundObstet Gynecol 1998; 5: 339-345.

30 Axt-Fliedner R. Second trimester uterine arteryDoppler ultrasound as a screening test foradverse pregnancy outcome. Clin Exp ObstetGynecol 2004; 1: 9-11.

31 Schwarze A, Nelles I, Krapp M, Friedrich M,Schmidt W, Diedrich K, Axt-Fliedner R.Doppler ultrasound of the uterine artery in the

prediction of severe complications during low-risk pregnancies. Arch Gynecol Obstet 2005; 1: 46-52.

32 Dugoff L, Lynch AM, Cioffi-Ragan D, HobbinsJC, Schultz LK, Malone FD, D’Alton ME. Firsttrimester uterine artery Doppler abnormalitiespredict subsequent intrauterine growthrestriction. Am J Obstet Gynecol 2005; 3 Pt 2:1208-1212.

33 Gomez O, Martinez JM, Figueras F, Del Rio M,Borobio V, Puerto B, Coll O, Cararach V, Vanrell JA. Uterine artery Doppler at 11-14weeks of gestation to screen for hypertensivedisorders and associated complications in anunselected population. Ultrasound ObstetGynecol 2005; 5: 490-494.

34 Herraiz I, Arbues J, Camano I, Gomez-MontesE, Graneras A, Galindo A. Application of a first-trimester prediction model for pre-eclampsiabased on uterine arteries and maternal historyin high-risk pregnancies. Prenat Diagn 2009;29: 1123-1129.

35 Coleman MA, McCowan LM, North RA. Mid-trimester uterine artery Doppler screening as a predictor of adverse pregnancy outcome inhigh-risk women. Ultrasound Obstet Gynecol2000; 1: 7-12.

36 Harrington K, Fayyad A, Thakur V, Aquilina J.The value of uterine artery Doppler in theprediction of uteroplacental complications inmultiparous women. Ultrasound Obstet Gynecol2004; 1: 50-55.

37 Axt-Fliedner R, Schwarze A, Nelles I, AltgassenC, Friedrich M, Schmidt W, Diedrich K. The value of uterine artery Doppler ultrasoundin the prediction of severe complications in a risk population. Arch Gynecol Obstet 2005; 1:53-58.

38 Cnossen JS, Morris RK, ter Riet G, Mol BW, vander Post JA, Coomarasamy A, Zwinderman AH,Robson SC, Bindels PJ, Kleijnen J, Khan KS. Useof uterine artery Doppler ultrasonography topredict pre-eclampsia and intrauterine growthrestriction: a systematic review and bivariablemeta-analysis. cmaj 2008; 6: 701-711.

39 Sciscione AC and Hayes EJ. Uterine arteryDoppler flow studies in obstetric practice. Am J Obstet Gynecol 2009; 2: 121-126.


40 Jurkovic D, Jauniaux E, Kurjak A, Hustin J,Campbell S, Nicolaides KH. Transvaginal colorDoppler assessment of the uteroplacentalcirculation in early pregnancy. Obstet Gynecol1991; 3: 365-369.

41 Jauniaux E, Zaidi J, Jurkovic D, Campbell S,Hustin J. Comparison of colour Dopplerfeatures and pathological findings in complicatedearly pregnancy. Hum Reprod 1994; 12: 2432-2437.

42 Locci M, Nazzaro G, De Placido G, Nazzaro A,Colacurci N, Montagnani S, Montemagno U.Correlation of Doppler and placentalimmunohistochemical features in normal andintrauterine growth-retarded fetuses. UltrasoundObstet Gynecol 1993; 4: 240-245.

43 Murakoshi T, Sekizuka N, Takakuwa K,Yoshizawa H, Tanaka K. Uterine and spiralartery flow velocity waveforms in pregnancy-induced hypertension and/or intrauterinegrowth retardation. Ultrasound Obstet Gynecol1996; 2: 122-128.

44 Matijevic R and Johnston T. In vivo assessmentof failed trophoblastic invasion of the spiralarteries in pre-eclampsia. Br J Obstet Gynaecol1999; 1: 78-82.

45 Hung JH, Ng HT, Pan YP, Yang MJ, Shu LP. Color Doppler ultrasound, pregnancy-inducedhypertension and small-for-gestational-agefetuses. Int J Gynaecol Obstet 1997; 1: 3-11.

46 Myatt L and Miodovnik M. Prediction of pre-eclampsia. Semin Perinatol 1999; 1: 45-57.

47 Conde-Agudelo A, Villar J, Lindheimer M.World Health Organization systematic reviewof screening tests for preeclampsia. ObstetGynecol 2004; 6: 1367-1391.

48 Masse J, Giguere Y, Kharfi A, Girouard J, ForestJC. Pathophysiology and maternal biologicmarkers of preeclampsia. Endocrine 2002; 1: 113-125.

49 Spencer K, Cowans NJ, Chefetz I, Tal J, Meiri H.First-trimester maternal serum pp-13, papp-a andsecond-trimester uterine artery Dopplerpulsatility index as markers of pre-eclampsia.Ultrasound Obstet Gynecol 2007; 2: 128-134.

50 Diab AE, El Behery MM, Ebrahiem MA, Shehata AE. Angiogenic factors for theprediction of pre-eclampsia in women with

abnormal midtrimester uterine artery Doppler velocimetry. Int J Gynaecol Obstet 2008;2: 146-151.

51 Poon LC, Kametas NA, Maiz N, Akolekar R,Nicolaides KH. First-trimester prediction ofhypertensive disorders in pregnancy. Hypertension2009; 5: 812-818.

52 Rizzo G, Capponi A, Cavicchioni O, Vendola M,Arduini D. First trimester uterine Doppler and three-dimensional ultrasound placentalvolume calculation in predicting pre-eclampsia. Eur J Obstet Gynecol Reprod Biol 2008; 2: 147-151.

53 De Vivo A, Baviera G, Giordano D, Todarello G,Corrado F, D’anna R. Endoglin, PlGF and sFlt-1as markers for predicting pre-eclampsia. ActaObstet Gynecol Scand 2008; 8: 837-842.

54 Lim JH, Kim SY, Park SY, Yang JH, Kim MY, Ryu HM. Effective prediction of preeclampsia by a combined ratio of angiogenesis-relatedfactors. Obstet Gynecol 2008; 6: 1403-1409.

55 Pihl K, Larsen T, Krebs L, Christiansen M. Firsttrimester maternal serum papp-a, beta-hCG andadam12 in prediction of small-for-gestational-age fetuses. Prenat Diagn 2008; 12: 1131-1135.

56 Semczuk-Sikora A, Krzyzanowski A, StachowiczN, Robak J, Kraczkowski J, Kwiatek M, SemczukM. [Maternal serum concentration of angiogenicfactors: pigf, vegf and vegfr-1 and placentalvolume in pregnancies complicated byintrauterine growth restriction]. Ginekol Pol 2007;10: 783-786.

57 Pilalis A, Souka AP, Antsaklis P, Daskalakis G,Papantoniou N, Mesogitis S, Antsaklis A.Screening for pre-eclampsia and fetal growthrestriction by uterine artery Doppler and papp-a

at 11-14 weeks’ gestation. Ultrasound ObstetGynecol 2007; 2: 135-140.

58 Acharya G and Morgan H. First-trimester, three-dimensional transvaginal ultrasound volumetryin normal pregnancies and spontaneousmiscarriages. Ultrasound Obstet Gynecol 2002; 6:575-579.

59 Cosmi E, Piazze JJ, Ruozi A, Anceschi MM, LaTR, Andrisani A, Litta P, Nardelli GB, AmbrosiniG. Structural-tridimensional study of yolk sac inpregnancies complicated by diabetes. J PerinatMed 2005; 2: 132-136.


60 Boito SME, Laudy JAM, Struijk PC, Stijnen T,Wladimiroff JW. Three-dimensional us

assessment of hepatic volume, headcircumference, and abdominal circumference inhealthy and growth-restricted fetuses. Radiology2002; 3: 661-665.

61 Gerards FA, Engels MA, Twisk JW, van Vugt JM.Normal fetal lung volume measured with three-dimensional ultrasound. Ultrasound ObstetGynecol 2006; 2: 134-144.

62 Yu C, Chang C, Chang F, Ko H, Chen H. Fetalrenal volume in normal gestation: a three-dimensional ultrasound study. Ultrasound MedBiol 2000; 8: 1253-1256.

63 Hafner E, Philipp T, Schuchter K, llinger-Paller B,Philipp K, Bauer P. Second-trimestermeasurements of placental volume by three-dimensional ultrasound to predict small-for-gestational-age infants. Ultrasound ObstetGynecol 1998; 2: 97-102.

64 Hafner E, Metzenbauer M, llinger-Paller B,Hoefinger D, Schuchter K, Sommer-Wagner H,Philipp K. Correlation of first trimester placentalvolume and second trimester uterine arteryDoppler flow. Placenta 2001; 8-9: 729-734.

65 Metzenbauer M, Hafner E, Hoefinger D,Schuchter K, Stangl G, Ogris E, Philipp K. Three-dimensional ultrasound measurement of theplacental volume in early pregnancy: methodand correlation with biochemical placentaparameters. Placenta 2001; 6: 602-605.

66 Metzenbauer M, Hafner E, Hoefinger D,Schuchter K, Philipp K. [Associations betweenbirth weight and placental volume in the firsttrimester]. Z Geburtshilfe Neonatol 2002; 4: 138-141.


68 Hafner E, Schuchter K, van LM, Metzenbauer M,llinger-Paller B, Philipp K. Three-dimensionalsonographic volumetry of the placenta and the fetus between weeks 15 and 17 of gestation.Ultrasound Obstet Gynecol 2001; 2: 116-120.


Reproducibility of three- dimensional ultrasound measurements of fetal and placental volume at gestationalages of 11-18 weeks

Koen L DeurlooMarieke S SpreeuwenbergMarijke Rekoert-HollanderJohn MG van Vugt

Journal of Clinical Ultrasound2007: 35 (3); 125-132

chapter 2

Abstract

objective: Determine the reproducibility of 3D ultrasound (3dus) measurements of fetaland placental volumes.

methods: We included 34 pregnant women between gestational weeks (gw) 11 – 18. Two operators independently acquired fetal and placental volumes using 3dus.Each volume was acquired twice and stored on disk for off-line analysis. Intra- and interobserver reproducibility was expressed in the intra- and interclasscorrelation coefficient (intra-cc and inter-cc). In addition, the 3dus volumes acquired by the first operator were calculated by the second and vice-versa toevaluate the effect of volume acquisition and caliper placement. A value > 0.75

was considered a good agreement.

results: Fetal and placental volume measurements were successful in 97% of all cases.Between gw 11-14 and 14-18 the median fetal volume was 20.8 (5.0 – 35.1) and51.7 (37.9 – 132.8) ml, the median placental volume was 71.3 (40.9 – 111.9) and 120.7 (94.2 – 273.7) ml. Bland-Altman plots were used for statistical analysis.The intraobserver reproducibility was good for fetus (intra-cc: 0.99; 0.99) andplacenta (intra-cc: 0.99; 0.98). Also, interobserver reproducibility was good for fetus (inter-cc 0.98) and placenta (inter-cc 0.98). In addition, regardless of the operator who acquired the volumes, the inter-cc remained good for bothfetus (inter-cc: 0.99; 0.99) and placenta (inter-cc: 0.97; 0.99).

conclusion: The reproducibility of fetal and placental volume measurements by 3dus

between gw 11 – 18 is good. In addition, individually chosen caliper placementand volume acquisition has no effect on the calculation of either volumes.

Introduction

In 1984 the first fetal volumes acquired by ultrasound were constructed by Brinkleyet al.1 Since then the development of three dimensional ultrasound (3dus) imagingassisted by computer technology made it possible to measure and calculate fetaland placental volumes quickly and accurately. Measuring and monitoring fetal andplacental volumes at different gestational ages may improve our understandingabout physiological and pathophysiological mechanisms in fetal and placentalgrowth. Fetal and placental volumes have been used in screening for fetuses withchromosomal abnormalities2-4, intrauterine growth restriction fetuses or preg-nancies complicated by preeclampsia5-7. There are reports in the literature of in-creased placental volumes preceding preeclampsia, decreased placental volumespreceding intrauterine growth restriction and decreased fetal volumes in fetuseswith chromosomal abnormalities.

Using 3dus for clinical purposes requires a critical evaluation of all aspects ofreproducibility, including the effect of operator selected caliper placement andvolume acquisition. Few studies have partially determined the reproducibility ofvolume measurement using 3dus.3,8 Whereas these studies have focused primarilyon intra- and interobserver variability, we consider the effect of volume acquisi-tion and individually traced circumferences – that is, the difference betweenmeasurements that are acquired by one operator and calculated by another – ofimportant additional value. The aim of this study was to determine all aspects ofintra- and interobserver reproducibility of fetal and placental volume measure-ments in different women at different gestational ages.

Patients and Methods

Our prospective study included women who presented to our institution’s obstetricdepartment for sonographic examination between 11 till 18 weeks of gestation.Only singleton pregnancies with a known last menstrual period and sonographi-cally confirmed gestational age were included. Our study was approved by theMedical Ethic Committee of VU University Medical Center. Informed consentwas obtained from all women after explanation of the study. Exclusion criteria weremultiple pregnancies and fetal anomalies. All measurements were performedtransabdominally using a Technos mx (Esaote; pie Medical, Maastricht, the Nether -lands) with a 3.5 MHz transabdominal annular-array probe with an externallyfixed 3D acquisition module and a magnetic tracking device.

21CHAPTE R 2 REPRODUCIBILITY OF THREE-DIMENSIONAL ULTRASOUND MEASUREMENTS OF FETAL AND PLACENTAL VOLUME AT GESTATIONAL AGES OF 11-18 WEEKS

Placenta and fetal volumes were acquired using methods developed by Hafner andcolleagues in 1998.5,9 The placental site was located using 2-dimensional real-timesonography, and a scanning plane parallel to the placental attachment area waschosen. The volume box was optimized relative to the size of the placenta andvolume scanning was performed. The volume was stored on magnetic opticdisks and calculated off-line to reduce examination time. For the actual placentalvolume calculation the axis of the narrowest placental depiction is used. A caliperwas used to trace the placental border and the circumference was recorded.This procedure was repeated in parallel sections with a variable distance with amaximum of 10 mm throughout the placental volume. At least 15 slices weremeasured for each placenta. The ultrasound equipment calculated the volumeusing the trapezoid volume formula V = ∑ (Aj + Aj-1) x (d j- d j-1), where A is thetraced area, d is the distance between the slices, j is the selected slice and j-1 isthe following slice.

The fetal volume was acquired and constructed using methods developed byHafner and colleagues.5,9 The fetus was depicted in a transverse view by 2-dimen -sional sonography. The volume box was again optimized relative to the size ofthe fetus, and volume scanning was performed. We repeated the scan when thefetus moved or showed breathing movements during measurement. Volumeswere stored and analyzed off-line. For the actual fetal volume calculation, slicesat right angles to the mid axial plane were used. A caliper was used to trace thefetal border and this procedure was repeated in at least 15 slices with a maximumdistance of 10 mm (figure 1). The measurement included head, neck, shoulders,and trunk; the extremities were excluded. Volume was calculated by the ultra-sound equipment using the trapezoid formula mentioned previously. From week18 onward, the size of the fetus and placenta usually exceeded the capacity ofthe volume box, thus total volume measurements could not be performed. Ourmeasurements were therefore made between 11 and 18 weeks’ gestation.

The 3dus scans were performed by two experienced operators. Placentaland fetal volumes were acquired twice by the first operator with repositioning ofthe probe after the first measurement. This was followed by the same volumeacquirement procedure by the second operator. Each operator was blinded forthe other’s volume acquirements. Total scanning time was approximately 5 to 6minutes. After examination of all patients, volumes were calculated off-line usingthe image analysis software of the ultrasound equipment. From this data, intra- andinterobserver variability expressed in intraclass correlation coefficient (intra-cc)and interclass correlation coefficient (inter-cc) were assessed. In addition, thevolumes acquired by the first operator were assessed by the second operator


and vice-versa to evaluate the effects of caliper placement and volume acquisition,expressed in inter-cc. A value of > 0.75 was considered a good consensus forthe intra- and inter-cc. Statistical analysis was performed using spss version 12.0(spss Inc., Chicago, il).

Intra-observer repeatability was expressed as the difference between twomeasurement results obtained by the same observer. Bland-Altman plots wereconstructed by plotting the difference between the 2 measurements against the mean of all measurements in order to assess the relationship between thedifference and the magnitude of the measurements. Limits of agreement (meandifference +/- 1.96 sd) were calculated as described by Bland en Altman10,11.

Systematic bias between the first and the second analysis was determined bycalculating the 95% ci for the mean difference. If zero lay between these intervals,no bias was assumed. Intra-observer repeatability was expressed as the intra-cc.

In the calculations for determining the inter-observer variance difference only


figure 1 – Multiplanar technique for the measurement of fetal volume. The grey lines in each plane represent the traced fetal circumference. Top left: saggital plane. Bottom left: transverseplane. Bottom right: coronal plane. Top right: Multiplanar 3-dimensional fetal volume.

the first measurement of an operator was used. Bland Altman plots were con-structed to assess the systematic bias between the 2 operators and the relation-ship between the difference between the 2 measurements and the magnitude ofthe measurements. The mean of all measurements was used as an estimate for themagnitude of the measurements. Bias between the two operators was deter-mined by calculating the 95% ci for the mean differences. Limits of agreementswere calculated as described by Bland and Altman. Inter-observer agreement wasexpressed as the inter-cc.

Bland Altman plots were also constructed to assess the systematic bias be-tween the measurements of both operators of a volume acquired by one of theoperators and the magnitude of the measurements. Inter-observer agreementwas also expressed as inter-cc. A value of > 0.75 was considered a good con-sensus for the intra-cc and inter-cc.

Results

Fetal and placental volume measurements were successfully performed in 33 of34 cases; one case was excluded due to poor image quality related to maternalobesity. The median maternal age was 34 and ranged from 21 to 43 years. Themedian gestational age was 14 weeks (range, 11 – 18 weeks). The mean and medianplacental and fetal volumes are shown in Table 1. The results of fetal and placentalvolumes by each operator are shown in Table 2. Both the intra- and interobserverdifferences are normally distributed according to the Kolmogorov-Smirnov test,which showed p-values > 0.05. Intraobserver reproducibility is shown in Table 3.

The Bland-Altman plots for the intra-observer variability are shown in Figures2 – 5. Interobserver variability is shown in Table 4. From the Bland-Altman plotsin Figure 6 and 7, it appears that there is no systematic bias between the 2 opera-tors. Intra-cc and inter-cc values were high for fetal and placental volumes. Thelimits of agreement were wide. There was no systematic bias between any pairedmeasurement results, nor a clear relationship between the differences and themagnitude of the measurements. Log-transformed scores did not improve theBland Altman plot.

From the Bland-Altman plots in Figures 8 to 11, it is clear that there was nodifference in placental and fetal volume measurements when one operator per-formed the acquisition with 3dus and the other operator traced the slices andcalculated the volumes. Table 5 shows high inter-cc values for both placental andfetal volumes, regardless of the acquiring or calculating operator.


table 1. Mean and median fetal and placental volume measurements (in ml) between gestational ages of 11 and 14 and 14 and 18 weeks.

Mean sd Median Range

Between 11-14 weeksFetus 21.5 8.16 20.8 5.00 - 35.1Placenta 73.9 22.2 71.3 40.9 - 112

Between 14-18 weeksFetus 67.9 29.9 51.7 37.9 - 133Placenta 131 47.2 121 94.2 - 274

table 2. Results of the First (a) and Second (b) Series of 33 measurements of Fetal and Placental Volumes (in ml) Obtained by Each of 2 Operators.

Operator Parameter Number of Mean sd median rangemeasurements

I Fetus (a) 33 145.2 33.0 136.7 35.1 - 142I Fetus (b) 33 146.0 33.2 137.5 34.7 - 144II Fetus (a) 33 145.4 32.0 135.8 35.2 - 122II Fetus (b) 33 145.0 31.0 138.6 35.1 - 122

I Placenta (a) 33 1102 46.9 194.0 39.1 - 271I Placenta (b) 33 1104 51.0 196.2 37.5 - 277II Placenta (a) 33 1105 46.4 1102 42.2 - 280II Placenta (b) 33 1103 45.1 193.9 40.1 - 267


table 3. Intraobserver Differences between the Results of the First and the Second Fetal and Placental Volume Measurement (in ml) for Each of the 2 Operators.

Intra-observer differenceOperator Parameter Mean Limits of agreement 95% ci Intra-cc

I Fetus -0.77 -6.92 , 5.38 -1.88 - 0.34 0.99II Fetus -0.34 -7.90 , 8.57 -1.15 - 1.83 0.99

I Placenta -2.17 -18.1 , 13.7 -5.05 - 0.70 0.99II Placenta -1.20 -15.6 , 18.0 -1.84 - 4.24 0.98

table 4. Interobserver Differences of Fetal and Placental Volumes (in ml) between Each Operator.

Intra-observer differenceParameter Mean Limits of agreement 95% ci Inter-cc

Fetus 0.14 -11.6 , 11.9 -1.99 - 2.26 0.98Placenta 2.47 -19.4 , 24.4 -1.49 - 6.44 0.98


27

figure 4 – Bland-Altman plot of the intraobserver differences for placental volumes for operator i.figure 5 – Bland-Altman plot of the intraobserver differences for placental volumes for operator ii.

figure 2 – Bland-Altman plot of the intraobserver differences for fetal volumes for operator i.figure 3 – Bland-Altman plot of the intraobserver differences for fetal volumes for operator ii.

figure 6 – Bland-Altman plots of the interobserver differences for fetal volumes between operators.figure 7 – Bland-Altman plot of the interobserver differences for placental volumes between operators.


figure 8 – Bland-Altman plot of the interobserver differences for fetal volumes acquired by operator i.

figure 9 – Bland-Altman plot of the interobserver differences for placental volumes acquired by operator i.

figure 10 – Bland-Altman plot of the interobserver differences for fetal volumes acquired by operator ii.

figure 11 – Bland-Altman plots of the interobserver differences for placental volumes acquiredby operator ii.

table 5. Interobserver Differences between Fetal and Placental Volumes Measurements Acquired by One Operators and Calculated by the Other, and Vice Versa.

Interobserver difference

Acquisition Calculation Compared Volume Number Mean sd Limits 95% Inter-

by by with of of ci cc

Operator Operator Calculation Images Agreement

of

Operator

I II I Fetus 33 1.02 3.79 -6.41 , -0.32 - 0.99

8.45 2.37

II I II Fetus 33 -0.17 2.54 -4.99 , -0.92 - 0.99

4.96 0.88

I II I Placenta 33 0.31 10.8 -20.9 , -3.54 - 0.97

21.5 4.15

II I II Placenta 33 -1.56 6.13 -13.6 , -3.37 - 0.99

10.5 0.61


Discussion

Sonographic measurement of fetal and placental volumes has been proposed asa tool for screening fetuses with chromosomal abnormalities2-4 or predicting in-trauterine growth restriction or pre-eclampsia5-7. Before using volume measure-ments in a clinical setting and assessing their predictive value, reproducibilityshould be tested.

The data reported in this study clearly show that the 3dus measurementsand calculation of fetal and placental volumes is highly reproducible both by thesame and by different operators. In addition, operator-selected caliper place-ment and volume acquisition has no significant effect on the calculation of eithervolume. Intraobserver agreement was very high for both fetal volume (0.99 and0.99) and placental volume (0.99 and 0.98). Additionally, interobserver scoreswere high for both fetal and placental volume (0.98 and 0.98, respectively). Theslightly higher variability of placental volumes is probably caused by less optimaldistinction of the placental margins from the uterine wall compared with themargins of the fetus. The latter is surrounded by amniotic fluid and therefore iseasily distinguished from its surroundings. Our data appear to indicate that thereis no significant effect of individual differences in caliper placements and volumeacquisition.

Few studies have partially assessed the reproducibility of fetal and placentalvolumes via 3dus. These studies primarily describe intra- and interobserver vari-ability, but do not mention the effect of individual differences in caliper placementand volume acquisition. Good inter- and intraobserver variabilities for placentalvolumes were first described in 1998 by Hafner et al.5 In a large pro spective studyby Wegrzyn et al.,3 417 fetuses and placentas were included. Bland-Altman plotsfor placental volume showed a mean difference between paired measurementsby same operator of 0.6 mL (limits of agreements [95%] – 6.1, – 7.4 mL) and amean difference between two operators of 0.7 mL (limits of agreement [95%] – 10.5, – 13.1 mL). These results are comparable to our results, although the meandifferences found by these authors are slightly smaller. Interobserver variability forfetal volumes was previously described by Aviram et al.12 in a study of 36 fetuses;no significant differences were found between 2 operators. No exact numberswere reported in this study, however. Good intra- and interobserver variability wasalso mentioned by Hafner et al.9 in 2001. More recently, Falcon et al.8 assessedvariability of fetal volume measurements, showing a mean difference betweenpaired measurements by the same operator of – 0.87 mL (limit of agree ment[95%] – 2.31, – 4.05 mL) and a mean difference between paired measurements by


two operators of – 1.09 mL (limits of agreement [95%] – 5.49, – 3.32 mL). Thesefindings are comparable to our results. Wegrzyn et al. and Falcon et al. used dif-ferent ultrasound equipment for their measurements, and we believe the differ-ence in image quality compared to our ultrasound equipment is responsible forthe small differences in intra- and interobserver variabilities for placental volumes.Higher image quality can improve distinction between the margins of the placentaand the uterine wall, thereby decreasing interobserver variability.

The effect of having caliper placement and volume acquisition performed byone operator and volume calculation performed by another has not been de-scribed previously. Our results show that there is no significant effect of thatcombination on the calculated placental and fetal volumes. This finding is helpfulwhen implementing volume scanning in prenatal screening programs: because itreduces total examination time.


References

1 Brinkley JF, McCallum WD, Muramatsu SK, et al. Fetal weight estimation from lengths and volumes found by three-dimensionalultrasonic measurements. J Ultrasound Med.1984;3:163.

2 Falcon O, Peralta CF, Cavoretto P, et al. Fetaltrunk and head volume in chromosomallyabnormal fetuses at 11 + 0 to 13 + 6 weeks of gestation. Ultrasound Obstet.Gynecol.2005;26:517.

3 Wegrzyn P, Faro C, Falcon O, et al. Placentalvolume measured by three-dimensionalultrasound at 11 to 13 + 6 weeks of gestation:relation to chromosomal defects. UltrasoundObstet.Gynecol. 2005;26:28.

4 Metzenbauer M, Hafner E, Schuchter K, et al.First-trimester placental volume as a marker for chromosomal anomalies: preliminary resultsfrom an unselected population. UltrasoundObstet.Gynecol. 2002;19:240.

5 Hafner E, Philipp T, Schuchter K, et al. Second-trimester measurements of placental volume by three-dimensional ultrasound to predictsmall-for-gestational-age infants. UltrasoundObstet.Gynecol. 1998;12:97.

6 Wolf H, Oosting H, Treffers PE. A longitudinalstudy of the relationship between placental andfetal growth as measured by ultrasonography.Am.J Obstet.Gynecol. 1989;161:1140.

7 Wolf H, Oosting H, Treffers PE. Second-trimester placental volume measurement by ultrasound: prediction of fetal outcome. Am.J Obstet.Gynecol. 1989;160:121.

8 Falcon O, Peralta CF, Cavoretto P, et al. Fetal trunk and head volume measured bythree-dimensional ultrasound at 11 + 0 to 13 + 6weeks of gestation in chromosomally normalpregnancies. Ultrasound Obstet.Gynecol.2005;26:263.

9 Hafner E, Schuchter K, van Leeuwen M, et al.Three-dimensional sonographic volumetry ofthe placenta and the fetus between weeks 15and 17 of gestation. Ultrasound Obstet.Gynecol.2001;18:116.

10 Bland JM, Altman DG. Statistical methods forassessing agreement between two methods ofclinical measurement. Lancet 1986;1:307.

11 Altman DG, Bland JM. Measurement in Medicine:the Analysis of Measurement ComparisonStudies. The Statistician 1983;32:307.

12 Aviram R, Shpan DK, Markovitch O, et al.Three-dimensional first trimester fetalvolumetry: comparison with crown rumplength. Early Hum.Dev. 2004;80:1.


Clinical use of three-dimensional ultrasound fetal and placental volumes

Koen L DeurlooJochem W KistJohn MG van Vugt

Adapted fromFetal and Maternal Medicine Review2007. 18 (2) 145-179

chapter 3

Introduction

Although three-dimensional ultrasound (3dus) was already being used in the early1980’s1-3 and the first commercially available three-dimensional scanner was pre-sented in 1989 in Paris at the International Radiology Congress4, it has only been inthe recent decades that the technology has found its way into clinical applications.This rapid development is mainly due to the equally rapid development of com-puter processor technology, essential for three-dimensional ultrasound. Originallyused as a tool for creating a surface mode showing the face of the fetus and genderabnormalities, its application has become much broader nowadays. The techniqueis now becoming more important in obstetrics and gynecology, but also in manyother fields such as cardiology, surgical applications and urology. Three-dimen-sional ultrasound is an easy, quick and cheap imaging technique without any knownadverse effects on the mother or developing fetus as stated by the American Institute of Ultrasound in Medicine (aium). In addition, it is reported by various authors to be very accurate in assessing volumes5-9. Furthermore, it has beensuggested that both two- and three-dimensional ultrasound, especially in the 1st

and 2nd trimester, are of positive influence in initiating the bonding between

mother and infant10.Although the technique of 3dus is relatively new, its clinical importance has

already been established in prenatal diagnosis and is likely to continue to increase.This review will provide an overview of the current applications of three-dimen-sional ultrasound measurements of placental and fetal volumes in fetal medicine.

Fetal volume

Acquiring fetal volumes can improve our understanding of the physiology andpathophysiology of fetal growth. Brinkley was the first to conclude in 1982 that3dus head-to-trunk reconstructions were accurate measurements of total fetalvolume. He performed two-dimensional ultrasound (2dus) fetal head-to-trunkvolume measurements on intra uterine demised fetuses and compared those withtheir post partum volume, measured by water displacement1. The study revealeda very high correlation (r = 0.999) between 3dus volumes and fetal volume; al-though fetal volume was underestimated due to the absence of 3dus limb volumemeasurement (regression coefficient = 1.73). A few years later, Brinkley alsoconducted a research of 41 fetuses within 48 hours prior to delivery and usedmultiple length and volume variables to estimate birth weight. From this data,


Brinkley concludes that 3dus measurements would provide the best predictionof fetal weight and would enable monitoring of fetal growth in the future, takinginto account the rapid evolving computer technology in the early 80’s11. Various3D reconstruction and volumes estimations using different time intervals havebeen done since Brinkley’s research. All studies focused on first and early secondtrimester measurements, since the size of fetuses in the late second and thirdtrimester usually exceeds the volume box and is therefore impossible to per-form. Fetal volumes were estimated in week 7 until 10 of gestation by Blaas et al.12,in week 11-14 by Hull13 in week 15-17 by Hafner14 and in week 11-18 of gestation byDeurloo et al.15 In 2006, we performed a study to assess reproducibility of 3dus

placental and fetal volume measurements in week 11-18 of gestation, which showedgood reproducibility and might be used in a clinical setting15.

Another up-to-date development is published by Blaas et al., who evaluatedfetal whole-body volumes in the first trimester with addition of fetal limbs16.Since fetal limbs represent a significant proportion of the fetal body volume, it isnot surprisingly that volume estimations are more accurate when these are in-cluded. However, the fetal volume needs to be segmented into different partswhich should be measured separately and the sum these parts will be the estimateof the total fetal volume. This requires advanced ultrasound equipment and ismore time consuming (approximately 20-30 minutes) than head-to-trunk volumemeasurements (approximately 5-10 minutes). Although limbs account for almost10% of the total volume, it is yet to be determined if these extensive measure-ments are a useful addition and have clinical relevance in detecting abnormalitiesin fetal growth and size.

In 2004, first trimester fetal volume normograms were created by Aviram et al.17

They studied 72 fetuses and found a high correlation between first trimestercrown-rump length and 3dus estimated fetal volumes (r = 0.94). A referencetable was constructed for week 6 to 12 which can serve as a tool for diagnosis ofearly pregnancy failure. Falcon performed several studies on fetal volumes to es-tablish the relationship between normal and aneuploid pregnancies in the firsttrimester18-20. He also created normograms for fetal head-trunk volumes basedon 417 fetuses, which he used for early identification of growth restriction in ane-uploid fetuses. They found about 40-45% smaller head-trunk volumes in fetuswith trisomy 13, 18 and triploidy, and 10-15% in Turner syndrome and trisomy 21

compared to normal fetuses and suggested that 3dus head-trunk volumes hasmore predictive value than crl measurements. In addition, head-to-trunk ratioswere used in normal and abnormal fetuses to differentiate between symmetrical

35CHAPTE R 3 CLINICAL USE OF THREE-DIMENSIONAL ULTRASOUND FETAL AND PLACENTAL VOLUMES

or asymmetrical growth restriction. Symmetrical growth restriction (equally af-fected trunk and head volumes) was found in trisomy 21 and Turner syndrome, incontrast to the asymmetrical growth restriction (trunk more affected than head) intrisomy 13, 18 and triploidy. The latter is possibly caused by placental insufficiencydue to placental maldevelopment and consequently redistribution in the fetal cir-culation.

Placenta

Measuring placental size and growth throughout gestation may differentiate be-tween physiology and pathophysiology of placenta growth. In the early eighties,Geirsson et al. studied the use of measuring placental volumes in normal and ab-normal pregnancies21,22. However, measurements were based on 2dus measure-ments and required complicated and time consuming computations. Since theintroduction of 3dus placental volumes can be acquired quickly and easily.

Reproducibility of measuring placental volumes has been examined thoroughlyand is found to be very good between different observers15, 23-26. Also, Metzen-bauer et al. showed that there seems a reasonable association between the firsttrimester placental quotient (placenta volume divided by crown-rump length) andbirth weight percentile in a group of 1476 pregnancies dived into, below the 10th,10

th to 50th, 50

th to 90th and above the 90

th percentile27. However, placenta vol-ume measurements are limited by the size of the volume box of the ultrasoundequipment and therefore can only be accurately performed before 18 weeks ofpregnancy15.

Placental volumes have been studied in their relationship to chromosomalabnormalities, hypertensive disorders in pregnancy and small-for-gestation fetuses.Wegrzyn et al. and Metzenbauer et al. studied placental volumes in chromosomalabnormal fetuses in the first trimester25,28. Metzenbauer et al. evaluated 17 fetuseswith a major chromosomal defect and used the placenta quotient to comparenormal and abnormal fetuses and found significantly lower median placenta quo-tients in the chromosomally abnormal group (abnormal versus normal: 0.67

versus 0.98). He suggested that this placenta quotient could be used as an im-provement of first-trimester screening for chromosomal anomalies. A few yearslater, Wegrzyn et al. evaluated 500 fetuses including 83 major chromosomal abnormalities and found only lower placenta volumes in trisomy 13 and 18. In tri-somy 21 and Turner syndrome placental volumes were not significantly differentcompared to normal fetuses.


He suggests that the combination of maternal age with fetal nt and maternalserum free beta-hCG and papp-a has a good detection rate of more than 90%.Since there is significant correlation between maternal serum free beta-hCG andpapp-a and placental volume, there is not much additional value to be expectedfrom additional placental volume measurements29. Therefore, Wegrzyn concludesthat first trimester measurement of placental volumes is not likely to improvedetection rate of chromosomal abnormalities.

Wegrzyn also evaluated the placental volumes in twin and triplet pregnancies inthe first trimester and constructed normograms. Placental volumes were dividedby the number of fetuses are respectively 83% and 76% of the placental volumesin singletons30. These percentages in the first trimester are the closely related tothe percentages when comparing birth weight and placental weight in twins ortriplets versus singleton pregnancies. The clinical relevance of these resultsshould be examined in the future, since there were no multiple pregnancies withadverse fetal or maternal outcome included.

Placenta volumes have also been studied to elucidate the relationship betweenplacenta volumes and alpha-thalassemia. Chen et al. suspected larger placentavolumes in fetuses affected by alpha-thalassemia due to fetal anemia and hypoxiaand subsequent placentomegaly26, but was unable to find significant differenceswith fetuses without alpha-thalassemia.

Most authors who examined placental volumes focused on their relationshipwith hypertensive disorders or small-for-gestation or growth restricted fetuses.Hafner et al. was the first to evaluate this relationship in week 16 to 23 of gestationand discovered decreased placenta volumes in sga fetuses (birth weight below10

th percentile). Unfortunately, predictive values were low with a sensitivity of52.5% and a specificity of 82.5%. In 2001, they evaluated placenta volumes inweek 15 to 17 of gestation and a quotient of fetal and placenta volume was devel-oped to relate placental size to fetal size and to be able to compare differentpregnancies more easily14. They suggest that the quotient of fetal and placentavolumes might assist in early detection of high risk pregnancies by the finding of alarge fetus with a small placenta. These findings are supported by the found rela-tionship between second trimester Doppler ultrasound in the uterine arteriesand placental size: placentas of women with high resistance uterine perfusion inthe second trimester are already remarkably small in the first trimester. There-fore placental volume measurements might be an efficient method for early iden-tification of impaired trophoblast invasion and subsequent hypertensive disordersin pregnancy24. Hafner et al. also investigated placental growth in normal preg-


nancies and pregnancies complicated by sga-fetuses or preeclampsia, but theyconclude that placental growth between week 12 and 22 is too heterogeneous tobe used as a clinical tool, although it might provide an insight in placental physiol-ogy and pathophysiology31.

Three-dimensional ultrasound can also be used to visualize placenta ab nor-malities, such as chorioangioma, hematoma, placental cysts or subplacental fibromas32. These abnormalities are already easily imaged by 2dus, but 3dus canenhance their visualization, increase their identification and act as a supplementto 2dus.

In conclusion

Three-dimensional fetal and placental volume measurements are easy, accurateand highly reproducible. The measurements are limited by gestational age, sincelate second trimester volume scans usually exceed the volume box. Clinical usefulinformation about aneuploid fetuses can be obtained in the first and early secondtrimester, although literature reports no increase in predictive values of currentscreening tests. First trimester measurements of fetal and placental volumes inprediction of hypertensive disorders or isolated iugr have not been publishedyet, and therefore their role in early identification remains yet to be determined.However, second-trimester placental volume measurements seem promising inearly identification of pregnancies complicated by growth restricted fetuses orhypertensive disorders.


References

1 Brinkley JF, McCallum WD, Muramatsu SK, Liu DY. Fetal weight estimation from ultrasonicthree-dimensional head and trunkreconstructions: evaluation in vitro. Am J ObstetGynecol 1982; 6: 715-721.

2 Baba K, Satoh K, Sakamoto S, Okai T, Ishii S.Development of an ultrasonic system for three-dimensional reconstruction of the fetus.J Perinat Med 1989; 1: 19-24.

3 Fredfeldt KE, Holm HH, Pedersen JF. Three-dimensional ultrasonic scanning. Acta Radiol Diagn (Stockh) 1984; 3: 237-241.

4 Maymon R, Herman A, Ariely S, Dreazen E,Buckovsky I, Weinraub Z. Three-dimensionalvaginal sonography in obstetrics and gynae -cology. Hum Reprod Update 2000; 5: 475-484.

5 Farrell T, Leslie JR, Chien PF, Agustsson P. The reliability and validity of three dimensionalultrasound volumetric measurements using an in vitro balloon and in vivo uterine model. bjog

2001; 6: 573-582.6 Gilja OH, Thune N, Matre K, Hausken T,

Odegaard S, Berstad A. In vitro evaluation ofthree-dimensional ultrasonography in volumeestimation of abdominal organs. Ultrasound MedBiol 1994; 2: 157-165.

7 King DL, King DL, Jr., Shao MY. Evaluation of in vitro measurement accuracy of a three-dimensional ultrasound scanner. J UltrasoundMed 1991; 2: 77-82.

8 Riccabona M, Nelson TR, Pretorius DH,Davidson TE. Distance and volumemeasurement using three-dimensionalultrasonography. J Ultrasound Med 1995; 12: 881-886.

9 Riccabona M, Nelson TR, Pretorius DH. Three-dimensional ultrasound: accuracy ofdistance and volume measurements. UltrasoundObstet Gynecol 1996; 6: 429-434.

10 Sedgmen B, McMahon C, Cairns D, Benzie RJ,Woodfield RL. The impact of two-dimensionalversus three-dimensional ultrasound exposureon maternal-fetal attachment and maternalhealth behavior in pregnancy. Ultrasound ObstetGynecol 2006; 3: 245-251.

11 Brinkley JF, McCallum WD, Muramatsu SK, Liu DY. Fetal weight estimation from lengths andvolumes found by three-dimensional ultrasonicmeasurements. J Ultrasound Med 1984; 4: 163-168.

12 Blaas HG, Eik-Nes SH, Berg S, Torp H. In-vivothree-dimensional ultrasound reconstructionsof embryos and early fetuses. Lancet 1998; 9135:1182-1186.

13 Hull AD, James G, Salerno CC, Nelson T,Pretorius DH. Three-dimensional ultrasono -graphy and assessment of the first-trimesterfetus. Journal of Ultrasound in Medicine 2001; 4:287-293.

14 Hafner E, Schuchter K, van LM, Metzenbauer M,llinger-Paller B, Philipp K. Three-dimensionalsonographic volumetry of the placenta and the fetus between weeks 15 and 17 of gestation.Ultrasound Obstet Gynecol 2001; 2: 116-120.

15 Deurloo KL, Spreeuwenberg MD, Rekoert-Hollander M, van Vugt JMG. Reproducibility of three-dimensional ultrasound measurementsof fetal and placental volume at gestational ages of 11-18 weeks. J Clin Ultrasound 2007; 35:

125-132.

16 Blaas HG, Taipale P, Torp H, Eik-Nes SH. Three-dimensional ultrasound volumecalculations of human embryos and youngfetuses: a study on the volumetry of compoundstructures and its reproducibility. UltrasoundObstet Gynecol 2006; 6: 640-646.

17 Aviram R, Shpan DK, Markovitch O, Fishman A,Tepper R. Three-dimensional first trimesterfetal volumetry: comparison with crown rumplength. Early Hum Dev 2004; 1: 1-5.

18 Falcon O, Cavoretto P, Peralta CF, Csapo B,Nicolaides KH. Fetal head-to-trunk volume ratioin chromosomally abnormal fetuses at 11 + 0 to13 + 6 weeks of gestation. Ultrasound ObstetGynecol 2005; 7: 755-760.

19 Falcon O, Peralta CF, Cavoretto P, Auer M,Nicolaides KH. Fetal trunk and head volume in chromosomally abnormal fetuses at 11+0 to13+6 weeks of gestation. Ultrasound ObstetGynecol 2005; 5: 517-520.

20 Falcon O, Peralta CF, Cavoretto P, Faiola S,Nicolaides KH. Fetal trunk and head volumemeasured by three-dimensional ultrasound at 11 + 0 to 13 + 6 weeks of gestation in


chromosomally normal pregnancies. UltrasoundObstet Gynecol 2005; 3: 263-266.

21 Geirsson RT. Intrauterine volume in pregnancy.Acta Obstet Gynecol Scand Suppl 1986; 1-74.

22 Geirsson RT, Ogston SA, Patel NB, Christie AD.Growth of total intrauterine, intra-amniotic andplacental volume in normal singleton pregnancymeasured by ultrasound. Br J Obstet Gynaecol1985; 1: 46-53.

23 Hafner E, Philipp T, Schuchter K, llinger-Paller B,Philipp K, Bauer P. Second-trimestermeasurements of placental volume by three-dimensional ultrasound to predict small-for-gestational-age infants. UltrasoundObstet Gynecol 1998; 2: 97-102.

24. Hafner E, Metzenbauer M, llinger-Paller B,Hoefinger D, Schuchter K, Sommer-Wagner H,Philipp K. Correlation of first trimester placentalvolume and second trimester uterine arteryDoppler flow. Placenta 2001; 8-9: 729-734.

25 Wegrzyn P, Faro C, Falcon O, Peralta CF,Nicolaides KH. Placental volume measured bythree-dimensional ultrasound at 11 to 13 + 6weeks of gestation: relation to chromosomaldefects. Ultrasound Obstet Gynecol 2005; 1: 28-32.

26 Chen M, Leung KY, Lee CP, Tang MH, Ho PC.Placental volume measured by three-dimensional ultrasound in the prediction of fetalalpha(0)-thalassemia: a preliminary report.Ultrasound Obstet Gynecol 2006;


28 Metzenbauer M, Hafner E, Schuchter K, Philipp K. First-trimester placental volume as amarker for chromosomal anomalies: preliminaryresults from an unselected population.Ultrasound Obstet Gynecol 2002; 3: 240-242.

29 Metzenbauer M, Hafner E, Hoefinger D,Schuchter K, Stangl G, Ogris E, Philipp K. Three-dimensional ultrasound measurement of the placental volume in early pregnancy:method and correlation with biochemicalplacenta parameters. Placenta 2001; 6: 602-605.

30 Wegrzyn P, Fabio C, Peralta A, Faro C,Borenstein M, Nicolaides KH. Placental volume

in twin and triplet pregnancies measured bythree-dimensional ultrasound at 11 + 0 to 13 + 6weeks of gestation. Ultrasound Obstet Gynecol2006; 6: 647-651.

31 Hafner E, Metzenbauer M, Hofinger D, MunkelM, Gassner R, Schuchter K, llinger-Paller B,Philipp K. Placental growth from the first to thesecond trimester of pregnancy in sga-foetusesand pre-eclamptic pregnancies compared tonormal foetuses. Placenta 2003; 4: 336-342.

32 Hata T, Kanenishi K, Inubashiri E, Tanaka H,Senoh D, Manabe A, Miyake K, Kondoh S, Hata T. Three-Dimensional SonographicFeatures of Placental Abnormalities. GynecolObstet Invest 2004; 2: 61-65.


Longitudinal Doppler measurements of spiral arteryblood flow in relation to uterine artery blood flow

Koen L DeurlooAntoinette C BolteJos WR TwiskJohn MG van Vugt

Journal of Ultrasound in Medicine2009; 28: 1623-1628

chapter 4

Abstract

objective:

The purpose of this study was to examine the longitudinal relationship betweenDoppler flow velocity waveforms of the spiral artery (ua) and uterine artery (sa)in pregnant women.

methods:

Ninety-seven primigravidas with uncomplicated singleton pregnancies wereanalyzed. Spiral artery (sa) and combined ua velocity waveforms were assessedby transabdominal ultrasound color Doppler at gestational ages of 11 through 13,14 through 17, and 18 through 24; each measurement was performed twice. Thepulsatility index (pi) was calculated for the left and right ua, and the results wereaveraged as a combined ua. In addition, the presence of ua bilateral notchingwas reported. Bland-Altman plots and generalized estimating equations wereused to assess intraobserver variability and the longitudinal relationship betweensa and ua blood velocities.

results:

A total of 284 ua and 263 sa Doppler flow measurements were analyzed.Intraobserver variability rates for the sa and the ua were 0.54 and 0.90,respectively. Results showed a continuous decrease of the mean pi in sa and ua

with increasing gestational age. Uterine artery bilateral notching was reported in 35%, 9% and 3% of the cases at gestational age 11 through 13, 14 through 17,and 18 through 24. Generalized estimating equation analysis showed a significantcorrelation (r = 0.41) between the sa and ua (p < 0.0001).

conclusion: Uterine artery Doppler measurements in early pregnancy seem to accuratelyreflect peripheral resistance of sa’s. Furthermore, trophoblastic invasion seemsa continuous process in the first half of pregnancy, in which early ua bilateralnotching is a physiological event.

Introduction

Uterine arteries (ua’s) are the main arteries for blood supply of the pregnantuterus. After branching from the hypogastric artery, they further divide into basaland radial arteries and, finally, into spiral arteries (sa’s). In the first half of preg-nancy during trophoblastic invasion, the sa’s are remodelled into dilated inelasticvessels capable of supplying large amounts of blood to the placenta and the devel -oping fetus. This physiologic process is only partly understood but is responsiblefor the development of a functional uteroplacental circulation.1,2

Impaired remodeling due to shallow trophoblastic invasion is associated withmaintenance of high- resistance low-flow sa’s as comparable to the nonpregnantuterus and related to possible subsequent development of intrauterine growthrestriction (iugr) and hypertensive disorders, such as preeclampsia (pe).2-5

Uterine artery Doppler measurements are thought to reflect the impedanceto flow in sa’s during trophoblastic invasion. The resistive index (ri) and pulsatilityindex (pi) as well as notching of the ua have been used alone or in combinationwith other maternal markers for early identification of pregnancies complicatedby preeclampsia, pregnancy-induced hypertension or iugr in numerous reports.6-9

The combination of high resistance and the persistence of bilateral notching untilweek 24 of gestation has been reported as a predictive tool for adverse pregnancyoutcome.10-15 Overall, these reports showed high negative predictive values butmoderate positive predictive values.

The ua’s and sa’s are presumed to be closely related, and first- or second-trimester sa flow velocity waveforms are quickly obtained by sonography.16-18

Therefore, several authors have examined the usefulness of sa Doppler meas-urements for screening purposes in complicated pregnancies; however, theirpredictive values remain questionable, ranging from good to poor.18-20

While we were studying the sa Doppler measurements for prediction of com -plicated pregnancies, we noticed that there were no studies available that confirmthe assumed relationship between sa and ua Doppler measurements. Therefore,we performed an analysis of the uncomplicated pregnancies in our study group toconfirm this relationship by using longitudinal data, and we assessed intraobserverreliability.

43CHAPTE R 4 LONGITUDINAL DOPPLER MEASUREMENTS OF SPIRAL ARTERY BLOOD FLOW IN RELATION TO UTERINE ARTERY BLOOD FLOW

Material and Methods

Women undergoing sonographic evaluations in our obstetric department beforeweek 11 of gestation were asked to participate in a prospective clinical study analyz -ing hypertensive disorders in pregnancy. To be included in the analysis, womenhad to be nulliparous with singleton pregnancies, a known last menstrual periodand a sonographically confirmed gestational age. Exclusion criteria were the pres-ence of fetal anomalies. In addition, women with complicated pregnancies suchas hypertension, preeclampsia, hellp (hemolysis, elevated liver enzyme levels, andlow platelets), and iugr were excluded from this analysis.

The Medical Ethics Committee of the VU University Medical Center approvedour study. Written informed consent was obtained from all women after expla-nation of the study. All measurements were performed transabdominally with anEsaote Technos mx Ultrasound system (pie Medical Imaging bv, Maastricht, theNetherlands) and a 3.5 MHz transabdominal annular array probe.

A single examiner performed all measurements in a quiet room with the patientin a supine position. Measurements were performed when uterine contractionswere absent and on 3 different gestational ages: 11 through 13, 14 through 17, and18 through 24 weeks. Each measurement was performed in a similar way: after afetal anatomy survey, fetal biometric measurements and amniotic fluid amountwere estimated and recorded. Then we used Color Doppler Sonography to detectboth ua’s and 2 different sa’s.

For detection of the sa’s, we used methods described by Murakoshi and co-workers19: we obtained flow velocity waveforms from the color-coded vasculararea just beneath the placenta. Spiral arteries were mainly identified by theircharacteristic waveform pattern but also discriminated from the radial and basalarteries by their localization just beneath the placenta (figure 1). When clear colorflow imaging was obtained, the Doppler gate was adjusted to the size of the ves-sel, usually 1 mm without changing the angle between the Doppler beam and thevessel because the direction of the flow is very difficult to determine. Wall filterswere set at a minimal level to detect the slowest blood flow. Waveforms were


figure 1 – Tracing of spiral artery blood flow.

sampled from 2 different sites beneath the central region of the placenta, at least3 cm away from the placental margin, avoiding insonation of sonographically de-tectable pathologic lesions such as infarction or clarification. Because there aredifferences in impedance indices obtained in the central region compared to theperipheral region of the placenta, we chose to perform all measurements withinthe central region of the placenta.

For detection of the ua, we used methods described by Jurkovic et al.16: bothua’s were identified on Color Doppler imaging, and sampling was performed oneach side of the ua at approximately 1 cm from the crossing points of the exter-nal iliac. The Doppler gate was adjusted to the size of the vessel, and the anglebetween beam and vessel was less than 10° to eliminate angle-dependent errors.Wall filters were set at a minimal level to detect the slowest blood flow, and at least3 sequential good waveforms were obtained. The diastolic notch was defined asa clearly defined upturn (angle of ≤90°) of the flow velocity waveform at the be-ginning of diastole (after the deceleration slope of the primary wave), which waspresent in all the waveforms.

Each sa and ua was measured twice with repositioning of the probe after thefirst measurement. Flow velocity waveforms were stored and analyzed online.Pulsatility indices were calculated immediately.

The course and outcome of the pregnancy were obtained by written follow-up and examination of the medical records. We recorded all complications duringpregnancy, the mode of delivery, the neonatal sex, the birth weight, and perinatalcomplications.

Data analysis

Spiral artery and ua flow velocities were analyzed by the pi: (peak systolic flow –end-diastolic flow) / mean total flow. In sa’s, the averaged pi’s values of 2 differentsites were used to calculate the mean peak systolic, end-diastolic and total flow.In ua’s, the averaged pi’s of the right and left artery were used.

Statistical analysis

Bland-Altman plots were used to assess intraobserver reliability for sa and ua

meas urements.21 Intra-observer reliability was expressed as an intraclass correla-tion coefficient (intra-cc).


To analyze the longitudinal relationship between combined sa and ua measure-ments, we used generalized estimating equations analysis in spss for Windowsversion 15.0 (spss Inc., Chicago, il). P < 0.05 was considered significant.

Results

A total of 136 pregnant women were included in the original study. Excluded werewomen with complicated pregnancies as described in “Material and Methods”(26 women) and parous women (13 women), resulting in 97 primigravidas withuncomplicated pregnancies. Mean maternal age was 32.3 (range 20-41) years, andthe mean bmi was 22.3 (range 17-28) kg/cm². Color Doppler measurements wereperformed successfully in most cases; ua’s were successfully measured in 98%(284/291) of all cases, sa’s in 90% (263/291) of all cases. Most common reasonsfor unsuccessful measurement were poor image quality due to maternal obesity,a posterior placenta and overlying fetal parts.

Analysis showed that the mean pi’s in spiral arteries and the right and left ua’sdecreased progressively with gestational age. In the sa’s, the mean pi decreasedfrom 0.54 (5th – 95

th percentiles, 0.30 – 0.77) to 0.42 (0.21 – 0.64) in weeks 11through 24 of gestation. In the same period, the mean pi in both left and right ua

decreased from 1.33 (0.10 – 2.54) to 1.01 (0.24 – 1.78; Table 1).Uterine artery bilateral notching was seen in 35% of all cases in gestational

week 11 through 13, decreasing with advancing gestational age to 9% and 3% ingestational weeks 14 through 17 and gestational weeks 18 till 24, respectively(Table 2).

The intraobserver differences were normally distributed. Intraobserver relia-bility was moderate for sa Doppler measurement (intra-cc = 0.54) and good forua measurements (intra-cc = 0.90). There was no systematic bias or relation-ship between the differences and the magnitude of the measurements. The meandifference was 0.01. A Bland-Altman plot for intraobserver reliability for sa

measurements is shown in figure 2.Generalized estimating equation analysis showed a significant longitudinal cor-

relation between combined ua and sa blood flow (p< 0.0001). The correlationcoefficient between the sa and both ua’s was 0.41 (Figure 3). When the totalgroup was divided in presence or absence of ua bilateral notching, only in caseswithout ua bilateral notching did a significant correlation (r = 0.33, p < 0.0001)between sa and ua pi remain; in cases with bilateral notching, no significant cor-relation (r = 0.18; P = 0.261) was found.


table 1. Mean pi’s of the sa and Combined Left and Right ua’s at Different Gestational Ages.

Gestational Age, wk Mean sa (± 2 sd) Mean ua (± 2 sd)

11 – 13+6 0.54 (0.30 – 0.77) 1.33 (0.10 – 2.56)14 – 17+6 0.51 (0.25 – 0.77) 1.21 (0.30 – 2.12)18 – 24+0 0.42 (0.21 – 0.64) 1.01 (0.24 – 1.78)

table 2. Numbers and Percentages of Pregnant Women with Absence or Presence of ua Bilateral Notching at Different Gestational Ages.

ua Bilateral notchingGestational Age, wk Absence Presence total Presence, %

11 - 13+6 62 33 95 3514 - 17+6 87 9 96 918 - 24+0 94 3 97 3


figure 2 – Bland-Altman plot of the intra-observer reliability for Spiral Artery Pulsatility Index.

figure 3 – Scatterplot and correlation coefficient of spiral and uterine artery pulsatility indices.

Discussion

The ua’s are responsible for the main blood flow to the pregnant uterus. Theybranch into sa’s, which become high-flow low-resistance vessels during tro-phoblastic invasion due to not completely elucidated physiologic mechanisms.From early pregnancy, the development of high-flow and low- resistance sa’s isassumed to be reflected in the decreasing impedance to flow in the ua’s, the l -atter being widely used as an early second-trimester predictive tool for compli-cated pregnancies, but although highly specific, they are generally lackingsen sitivity.12-15,22,23 Therefore, several authors have focussed on the evaluation ofsa Doppler measurements instead of uterine arteries for predictive purposes.18-

20 Although generally assumed to be correlated, to our knowledge, the relation-ship between sa and ua Doppler measurements has not been reportedpreviously. Our results support the assumption that ua Doppler measurementsare a reflection of changes in the sa’s that occur during placentation.

Furthermore, our longitudinal study results show that the pi’s of the sa’s andua’s decrease progressively with advancing gestational age, our findings and ref-erence ranges are consistent with previous reported literature.2,19,20 These studiessuggest that trophoblastic invasion occurs continuously in the first and secondtrimesters of pregnancy, which is in concordance with histological studies onuteroplacental sites in uncomplicated pregnancies and in contrast to the ‘2-waveinvasion’ model by Pijnenborg.1,2

Our results show that sa and ua Doppler measurements are significantlycorrelated. A decrement in sa Doppler indices is seen when ua Doppler indicesdecrease. This confirms the suggestion that impedance to sa blood flow is reflect-ed in ua blood flow. However, correlation coefficients were moderate, whichwas probably due to the measurement of 2 sa’s in a large amount of sa’s (ap-proximately >100 arteries) in different state of trophoblastically induced modifi-cation. This was also reflected in the moderate intra-cc’s for the sa Dopplermeasurements.

Uterine artery bilateral notching in this uncomplicated group shows a highincidence (35%) in the first trimester, which decreases with increasing gestationalage to approximately 3% in late second trimester. We think this reflects normalplacental development in which high-resistance uterine vessels progress to low-resistance vessels due to sa modification by progressive invasion of trophoblasticcells. Therefore, because first-trimester bilateral notching seems a regular physi-ologic phenomenon in uncomplicated pregnancies, it is probably not a specificmarker for early detection of complicated pregnancies. In our study, a significant


correlation between ua and sa Doppler measurements remained in the groupwithout bilateral notching. The absence of a significant relationship in the groupwith bilateral notching is most likely due to small numbers in that group, especiallyin the late second trimester (Table 2).

To our knowledge, a study that extensively examined the relationship be-tween sa and ua Doppler measurements has not been reported previously, andon the basis of our results, we conclude that they are significantly correlated. Es-pecially since our results are gained from a prospective longitudinal setting, wethink that our data give a realistic view on this relationship.

In addition, because primigravidas have higher ua ri’s compared with parouswomen, we choose to solely include primigravidas in our analysis to avoid dilutingour primary data and increase the change of finding relevant significant correla-tions.24

However, we did find moderate intra-cc’s for sa Doppler measurementsand high intra-cc for ua Doppler measurements. Because the ua’s are 2 clearlydefined vessels, high intra-cc are easily achieved. On the contrary, sa’s are numer-ous, and because of regional difference in trophoblastic invasion and subsequentdifferences in arterial wall modification, Doppler measurements can thereforeresult in different pi and ri outcomes. Increasing the number of measurementswill minimize this issue but will cause lengthening of examination time, which canbe costly and uncomfortable for patients.

Compared with ua Doppler flow measurements, sa Doppler flow measure-ments are unlikely to have better predictive values because of broad differencesin individual sa flow capacity and velocity. Although significantly correlated withua’s, the individual differences in sa’s are wide between individual vessels, and itis more likely that ua Doppler measurements in combination with the absence orpersistence of bilateral notching in the early second trimester (gestational weeks18-24) are better predictive tools for adverse pregnancy outcome.

We conclude that in uncomplicated pregnancies, multiple ua Doppler meas-urements clearly reflect the decreasing impedance to flow in sa’s and thereforethe ongoing placental development. In addition, trophoblastic invasion seems acontinuous process in which first-trimester bilateral notching is probably a physi-ologic phenomenon. It seems unlikely that sa Doppler measurements will bemore specific predictive tools than ua Doppler measurements for earlier identi-fication of complicated pregnancies.


References

1 Pijnenborg R, Bland JM, Robertson WB, Brosens I. Uteroplacental arterial changesrelated to interstitial trophoblast migration in early human pregnancy. Placenta 1983; 4: 397-413.

2 Lyall F. Priming and remodelling of humanplacental bed spiral arteries during pregnancy– a review. Placenta 2005; 26 (suppl A) S31-S36.

3 Brosens JJ, Pijnenborg R, Brosens IA. The myometrial junctional zone spiral arteries in normal and abnormal pregnancies: a reviewof the literature. Am J Obstet Gynecol 2002; 187: 1416-1423.

4 Sheppard BL and Bonnar J. An ultrastructuralstudy of utero-placental spiral arteries inhypertensive and normotensive pregnancy andfetal growth retardation. Br J Obstet Gynaecol1981; 88: 695-705.

5 Pijnenborg R, Anthony J, Davey DA, Rees A,Tiltman A, Vercruysse L, van AA. Placental bedspiral arteries in the hypertensive disorders ofpregnancy. Br J Obstet Gynaecol 1991; 98: 648-655.

6 Crispi F, Llurba E, Dominguez C, Martin-GallanP, Cabero L, Gratacos E. Predictive value ofangiogenic factors and uterine artery Dopplerfor early- versus late-onset pre-eclampsia andintrauterine growth restriction. UltrasoundObstet Gynecol 2008; 31: 303-309.

7 Spencer K, Cowans NJ, Chefetz I, Tal J, Meiri H.First-trimester maternal serum pp-13, papp-a

and second-trimester uterine artery Dopplerpulsatility index as markers of pre-eclampsia.Ultrasound Obstet Gynecol 2007; 29: 128-134.

8 Spencer K, Cowans NJ, Chafetz I, Tal J,Kuhnreich I, Meiri H. Second-trimester uterineartery Doppler pulsatility index and maternalserum pp13 as markers of pre-eclampsia. PrenatDiagnos 2007; 27:258-263

9 Rizzo G, Capponi A, Cavicchioni O, Vendola M,Arduini D. First trimester uterine Doppler andthree-dimensional ultrasound placental volumecalculation in predicting pre-eclampsia. Eur JObstet Gynecol Reprod Biol 2007; 138: 147-151

10 Coleman MA, McCowan LM, North RA. Mid-trimester uterine artery Doppler screening as a predictor of adverse pregnancy outcome in

high-risk women. Ultrasound Obstet Gynecol2000; 15: 7-12.

11 Papageorghiou AT, Yu CK, Bindra R, Pandis G,Nicolaides KH. Multicenter screening for pre-eclampsia and fetal growth restriction bytransvaginal uterine artery Doppler at 23 weeksof gestation. Ultrasound Obstet Gynecol 2001; 18:441-449.

12 Zimmermann P, Eirio V, Koskinen J, Kujansuu E,Ranta T. Doppler assessment of the uterine anduteroplacental circulation in the second-trimester in pregnancies at high risk for pre-eclampsia and/or intrauterine growthretardation: comparison and correlationbetween different Doppler parameters.Ultrasound Obstet Gynecol 1997; 9: 330-338.

13 Parretti E, Mealli F, Magrini A, Cioni R, Mecacci F, La TP, Periti E, Scarselli G, Mello G.Cross-sectional and longitudinal evaluation ofuterine artery Doppler velocimetry for theprediction of pre-eclampsia in normotensivewomen with specific risk factors. UltrasoundObstet Gynecol 2003; 22: 160-165.

14 Irion O, Masse J, Forest JC, Moutquin JM.Prediction of pre-eclampsia, low birthweight for gestation and prematurity by uterine arteryblood flow velocity waveforms analysis in lowrisk nulliparous women. Br J Obstet Gynaecol1998; 105: 422-429

15 Chien PF, Arnott N, Gordon A, Owen P, KhanKS. How useful is uterine artery Doppler flowvelocimetry in the prediction of pre-eclampsia,intrauterine growth retardation and perinataldeath? An overview. bjog 2000; 107: 196-208.

16 Jurkovic D, Jauniaux E, Kurjak A, Hustin J,Campbell S, Nicolaides KH. Transvaginal colorDoppler assessment of the uteroplacentalcirculation in early pregnancy. Obstet Gynecol1991; 77: 365-369.

17 Hung JH, Ng HT, Pan YP, Yang MJ, Shu LP. ColorDoppler ultrasound of spiral arteries in normalsecond-trimester pregnancies. Zhonghua Yi XueZa Zhi (Taipei) 1997; 59: 289-294.

18 Deurloo KL, Spreeuwenberg MD, Bolte AC,Van Vugt JM. Color Doppler ultrasound of spiralartery blood flow for prediction of hypertensivedisorders and intra uterine growth restriction: a longitudinal study. Prenat Diagn 2007; 27: 1011-1016.


19 Murakoshi T, Sekizuka N, Takakuwa K,Yoshizawa H, Tanaka K. Uterine and spiralartery flow velocity waveforms in pregnancy-induced hypertension and/or intrauterinegrowth retardation. Ultrasound Obstet Gynecol1996; 7: 122-128.

20 Hung JH, Ng HT, Pan YP, Yang MJ, Shu LP. ColorDoppler ultrasound, pregnancy-inducedhypertension and small-for-gestational-agefetuses. Int J Gynaecol Obstet 1997; 56: 3-11.

21 Bland JM, Altman DG. Statistical methods forassessing agreement between two methods ofclinical measurement. Lancet 1986; 1:307-310.

22 Harrington K, Carpenter RG, Goldfrad C,Campbell S. Transvaginal Doppler ultrasound of the uteroplacental circulation in the earlyprediction of pre-eclampsia and intrauterinegrowth retardation. Br J Obstet Gynaecol 1997;104: 674-681.

23 Kurdi W, Campbell S, Aquilina J, England P,Harrington K. The role of color Dopplerimaging of the uterine arteries at 20 weeks’gestation in stratifying antenatal care. UltrasoundObstet Gynecol 1998; 12: 339-345.

24 Pilalis A, Souka AP, Antsaklis P, Basayiannis K,Benardis P, Haidopoulos D, Papantoniou N,Mesogitis S, Antsaklis A. Screening for pre-eclampsia and small for gestational age fetusesat the 11-14 weeks scan by uterine arteryDopplers. Acta Obstet Gynecol Scand 2007; 86:530-534.


Color Doppler ultrasound of spiral artery blood flow for prediction of hypertensive disorders and intra uterine growthrestriction: a longitudinal study

Koen L DeurlooMarieke D SpreeuwenbergAntoinette C BolteJohn MG van Vugt

Prenatal Diagnosis2007; 27: 1011-1016

chapter 5

Abstract

objective:

To construct reference ranges for spiral artery (sa) flow velocities and examine the possibility to predict intra uterine growth restricted (iugr) fetus,pregnancy-induced hypertension (pih) and/or preeclampsia (pe).

methods:

Spiral artery blood flow velocity measurements were performed using ColorDoppler between 11 to 13+6, 14 to 17+6 and 18+0 to 24 weeks of gestation,each measurement was performed twice. Spiral artery flow velocities wereanalyzed with multilevel modeling: individual regression curves were estimatedand combined to obtain the reference intervals for sa flow velocities in normal pregnancies. Mann-Whitney U tests was used to compare the deviationfrom expected flow velocity between normal and complicated pregnancies.

results:

One hundred and eight pregnancies were included; 10 pregnancies werecomplicated with iugr fetuses (<P10), 7 were complicated with pih and 4 pregnancies were complicated with preeclampsia. In the uncomplicatedpregnancies, systolic/diastolic (sd) ratios, resistance index (ri) and pulsatilityindex (pi) decreased progressively with advancing gestational age. The best fits for sd, ri and pi versus gestational age were linear regression equations.According to these equations, the mean predicted sd ratio decreased from 1.75

at 11 weeks of gestation (P5 – P95: 1.32 – 2.17) to 1.48 at 24 weeks of gestation(P5 – P95: 1.01 – 2.2). The predicted ri en pi showed similar decrement: mean ri from 0.44 (P5 – P95: 0.26 – 0.60) at 11 weeks of gestation to 0.34 (P5 – P95: 0.15 – 0.50) at 24 weeks of gestation, mean pi from 0.59 (P5 – P95: 0.31 – 0.75) at 11 weeks of gestation to 0.40 (P5 – P95: 0.11 – 0.66) at 24 weeks of gestation. The Mann-Whitney tests found no difference in ri, sd-ratio or pi

in pregnancies complicated by pe and iugr fetuses. However, significantly lowerri, pi and sd-ratio were found preclinically in pregnancies complicated by pih.

conclusion:

Spiral artery flow velocities decrease progressively with advancinggestational age. Reference ranges for sa blood velocities are wide andearly measurements of pi, ri or sd-ratios are not sensitive enough topredict preeclampsia, pih or iugr fetuses. The significantly lower sa

impedance values preceding pregnancies complicated by pih are likelycaused by a hyperdynamic circulation in women with preexistenthypertension.

Introduction

Hypertensive disorders occurring during pregnancy are complications with poorlyunderstood etiology, but it is generally believed that the impaired trophoblasticinvasion and the subsequent lack of modification of spiral arteries play an impor-tant role in the development of these complications.1 Histological studies showthat in uncomplicated pregnancies the invasion of the trophoblastic cells extendsto the decidual and myometrial segments, changing the spiral arteries to low- resistance vessels capable of supplying high amounts of blood to the placenta. Inhypertensive disorders, the trophoblastic invasion is usually restricted to the decid-ual segments leaving the myometrial segment of the spiral arteries unchanged andretaining their musculo-elastic wall.1-3 This results in vessels with high resistance,these could eventually lead to impaired uteroplacental perfusion and subsequenthypertensive disorders or intra uterine growth restriction (iugr).4-6

Non-invasive study of the spiral artery velocimetry can be done easily usingColor Doppler ultrasound. One of the first (transvaginal) Color Doppler meas-urements was performed in the early nineties by Jurkovic and coworkers, sincethen the technological advances in obstetric ultrasound made Color Dopplermeasurements of spiral arteries more precise and accessible for transabdominalapproach.7 Color Doppler velocimetry and reference ranges for spiral arteries inuncomplicated pregnancies have been described by the same authors, showingprogressively decreasing indices of impedance to flow with advancing gestationalage.7 The relationship between the blood flow velocimetry and pregnancies com -plicated by hypertensive disorders or iugr have also been described before byseveral authors with different results.8-10 However, these studies were based oncross-sectional or matched-pairs data. The current study is a prospective longi-tudinal study and we used multilevel analysis for statistical analysis. Therefore, ourobjective was to establish longitudinal reference ranges for serial measurementsof spiral artery blood flow velocities and then find their relationship with pre -eclampsia (pe), pregnancy induced hypertension (pih) and pregnancies compli-cated by iugr.

Methods

Women, who presented in our obstetric department for ultrasound evaluationbefore week 11 of gestation, were asked to participate in our prospective clinicaltrial. Since hypertensive disorders in pregnancy have low incidence, we included


only those women at risk for developing a complicated pregnancy, such as primi-gravidas or women with a history of a complicated pregnancy. In addition, theirpresent pregnancies had to be singleton with known last menstrual period andultrasound confirmed gestational age. Exclusion criteria were multiple pregnan-cies and/or fetal anomalies. The Medical Ethic Committee of VU UniversityMedical Center approved our study. Written informed consent was obtainedfrom all women after explanation of the study. All measurements were per-formed transabdominally using a Technos mx (Esaote – pie Medical, Maastricht,the Netherlands) with a 3.5 MHz transabdominal annular array probe with an in-tegral 3D program.

A single examiner performed all measurements in a quiet room with the patientin supine position. Measurements were performed when uterine contractionswere absent, and on three different occasions: one measurement between week11-13+6, one between 14-17+6 and one in 18-24 of gestation. Each measurementwas performed in a similar way: after fetal anatomy survey, the fetal biometryand amniotic fluid amount were estimated and recorded. Then we used ColorDoppler ultrasound to detect at least two spiral arteries. For detection of thespiral arteries we used methods described earlier by Murakoshi and coworkers:we obtained flow velocity waveforms from the color-coded vascular area justbeneath the placenta. Spiral arteries were mainly identified by their characteristicwaveform pattern (Figure 1) and also discriminated from the radial and basal ar-teries by their localization just beneath the placenta. When clear color flow imag-ing was obtained, the Doppler gate was adjusted to the size of the vessel, usually1 mm without changing the angle between the Doppler beam and the vessel sincethe direction of the flow is very difficult to determine. Wall filters were set at aminimal level to detect the slowest blood flow. Waveforms were sampled from atleast two different sites beneath the central region of the placenta, at least 3 cmaway from the placental margin, avoiding insonation of sonographically detectablepathological lesions such as infarction or clarification.10 Since there are significant

57CHAPTE R 5 COLOR DOPPLER ULTRASOUND OF SPIRAL ARTERY BLOOD FLOW FOR PREDICTION OFHYPERTENSIVE DISORDERS AND INTRA UTERINE GROWTH RESTRICTION: A LONGITUDINAL STUDY

figure 1 – Characteristic spiral artery waveform.

differences in impedance indices obtained in the central region compared to theperipheral region of the placenta, we choose to perform all measurements with-in the central region of the placenta. Each artery was measured twice with repo-sitioning of the probe after the first measurement. Flow velocity waveforms werestored and analyzed online.

The course and outcome of the pregnancy were obtained by written followup and examination of the medical notes. We recorded all complications duringpregnancy, mode of delivery, neonatal sex and birth weight and postnatal com-plications. Hypertensive disorder definitions were according to the Report of theNational High Blood Pressure Education Program Working Group on High BloodPressure in Pregnancy11: pih was defined as systolic blood pressure above 140

mmHg and/or diastolic blood pressure above 90 mmHg, in pe this is accompa-nied by urinary excretion of protein of at least 300 mg in a 24-h specimen.Growth restricted fetuses were defined as birth weight below the tenth percentilefor gestational age.

Data analysis

Spiral artery flow velocities were assessed by calculating the resistance index ri

(peak systolic flow – end-diastolic flow / peak systolic flow), the pulsatility index(pi) (peak systolic flow – end-diastolic flow / mean total flow) and the systolic-diastolic (sd)-ratio (peak systolic flow / end-diastolic flow). The averaged valuesof two different sites were used to calculate peak systolic, end-diastolic and meantotal flow.

Statistical analysis

To determine the influence of patient characteristics on complicated pregnancieswe used a one way anova with a Bonferroni post-hoc comparison for continu-ous variables (maternal age, primigravida (%), bmi, gestational age at delivery andbirth weight). Multilevel modeling was used to obtain the relationship betweenthe spiral artery Doppler parameters (sd-ratio, pi and ri) and gestational age foruncomplicated pregnancies. In our analysis, we defined a two–level hierarchy toconstruct the multilevel model (level 1 = measurement time, level 2 = subject).

The spiral artery Doppler parameters were used as dependent variables andgestational age as independent variable controlled for possible confounders. Cu-


bic and quadratic polynomials were used to obtain the best fitting curves in rela-tion to gestational age for each spiral artery Doppler parameter. For each spiralartery Doppler parameter uncomplicated pregnancy 95% reference ranges wereconstructed.

Multilevel analysis is a technique that takes into account that the same womenare repeatedly measured and it uses all the available data; irrespectively of thenumber of repeated measurements. This indicates that missing data are allowed.Furthermore, multilevel analysis is capable of dealing with irregularly time inter-vals.12 Multilevel analysis was performed with the statistical software MLwin, version2.01 (Multilevel project, Institute of education, London, uk).

To investigate whether spiral artery flow velocity could predict hypertensivedisorders during pregnancies, such as pe or pih or predict growth restricted fetuses(iugr), a second multilevel analysis was performed with data from both complicat-ed and uncomplicated pregnancies. In this analysis, gestational age and three dummyvariables, which indicated each complicated pregnancies (pe, iugr and pih) wereused in the analysis, together with the possible confounders.

Chi square tests for the significance of the regression coefficients were used toconclude whether complicated pregnancies predicted spiral artery flows. P-valuesof p< 0.05 were considered significant.

Results

All 108 pregnant women completed the study and follow up was received. Riskfactors in our population were listed in table 1. Uncomplicated pregnancy wasreported in 86 women. There was one fetal death at 35 + 3 weeks due to placentalabruption, but this pregnancy was not complicated by pih, pe or iugr and there-fore excluded. In the total group of included women, four women were diagnosedwith pe and seven with a pih. In addition, ten neonates had birth weights belowthe tenth percentile (iugr). The overall median maternal age was 32 (20-40) years;median body mass index was 22.3 (17.7-28.4) kg/cm2 at inclusion. No statisticaldifferences were found with respect to maternal age, number of primigravida orgestational age at delivery between all groups, but as expected, birth weight wassignificantly lower in pregnancies complicated by iugr and pe. Maternal bmi at in-clusion was significantly higher in pregnancies complicated by pih compared touncomplicated pregnancies (Table 2).

The intra-observer variation for the ri’s, the pi’s and the sd-ratio’s was 0.61,0.59 and 0.54, respectively, meaning fair reproducibility. Retrieving satisfying flow


table 1. Risk factors in 108 included women.

Risk factors Number

Primigravida 93 (81 %)History of pe 3 (2.6 %)History of iugr 2 (1.8 %)Preexistent ht 4 (3.7 %)Diabetes Mellitus type i 1 (0.9 %)

pe, preeclampsia; iugr, intra uterine growth restriction; ht, hypertension.

table 2. Patient characteristics. Data presented as number of cases (%) or mean ±sd.

ClassificationUncomplicated Preeclampsia pih iugr p-value

Number of patients 86 4 7 10 NSAge (yrs) 32.4 ±3.8 30.3 ±1.5 34.7 ±3.5 32.9 ±4.0 NPrimigravida (%) 75 (87) 4 (100) 4 (57) 9 (90) NSbmi (kg/cm²) 22.3 ±2.0 20.4 ±2.5 24.6 ±2.2* 21.5 ±2.0 p< 0.01ga at delivery (days) 279 ±12 267 ±12 274 ±11 270 ±12 NSBirth weight (gram) 3591 ±449 2716 ±511* 3447 ±464 2431 ±492* p< 0.01

* p < 0.01bmi, Body Mass Index at inclusion; ga, gestational age; pih, pregnancy induced hypertension;iugr, intra uterine growth restriction.


velocity waveforms of the spiral artery was possible in nearly all cases (over 99%)and usually within five minutes.

Multilevel modeling showed that in uncomplicated pregnancies pi, ri and sd ratiodecreased progressively with gestational age. Cubic and quadratic polynomials ofgestational age did not improve the fit, nor was there an effect for neonatal weight,fetal heart rate and gender. In all multilevel models, based on the -2log likelihoodtest result, a random intercept model was chosen. The mean predicted sd ratio,based on the multilevel model, decreased from 1.75 at 11 weeks of gestation (P5-P95: 1.32-2.17) to 1.48 at 24 weeks of gestation (P5-P95: 1.01-2.20). The predicted ri

and pi showed similar decrement: mean ri from 0.44 (P5-P95: 0.26-0.60) at 11 weeksof gestation to 0.34 (P5-P95:0.15-0.50) at 24 weeks of gestation, mean pi from0.59 (P5-P95: 0.31-0.75) at 11 weeks of gestation to 0.40 (P5-P95:0.11-0.66 at 24)weeks of gestation.

The results from the random intercept multilevel analysis with both uncom-plicated and complicated pregnancies showed for iugr complicated pregnancies nodifferences with uncomplicated pregnancies in spiral artery flow for pi (p = 0.72), ri

(p = 0.99) and sd-ratio (p = 0.95). Also for pe complicated pregnancies did not differ with normal pregnancies in pi (p = 0.65), ri (p = 0.59) and sd-ratio (p = 0.59).Pregnancies complicated by pih showed significant lower pi (p = 0.01), ri (p = 0.03)and sd-ratio flows (p = 0.02). (figures 2-4).


figure 2

Pulsatility index in pe, pih

and iugr-complicated pregnancies in week 11 – 24

of gestation plotted on uncomplicated pregnancyreference ranges (P5 – P95).


figure 3

Resistance index in pe, pih

and iugr-complicated pregnancies in week 11 – 24


figure 4

sd -Ratio in pe, pih andiugr-complicated pregnancies in week 11 – 24


Discussion

In uncomplicated pregnancies the trophoblastic invasion causes transformationof the musculo-elastic wall of the spiral arteries in the decidual and myometrialsegment. The trophoblastic cells transform these vessels into low resistance vesselsunresponsive to circulating pressor agents and capable of supplying large amountsof blood to the placenta.13 During the initial wave spiral arteries in the decidua arealtered; during the second wave, which is normally completed at 20-22 weeks ofgestation, spiral arteries in the myometrium are altered. In pregnancies compli-cated by hypertensive disorders the second wave of trophoblastic invasion is im-paired and mostly restricted to the decidual segment of the spiral arteries. Themyometrial segment of the spiral artery maintains its muscular layer and there-fore its high resistance. This may lead to impaired uteroplacental perfusion causingcomplications later in pregnancy.14

Our study results show that indices of impedance to flow of the spiral arter-ies decrease progressively with advancing gestational age and these findings areconsistent with previous reported literature.10,15 These studies suggest that tro-phoblastic invasion is occurring continuously throughout the first and secondtrimester of pregnancy. This is also in concordance with histological studies onplacentas from uncomplicated pregnancies.13 Our reference ranges for pi, ri andsd-ratio are closely comparable to previous reported results by several authors10,15,although our time interval of measurements was different: gestational week 11-24

in our study versus week 13-25 in Hung’s study and week 18-41 in the study ofMurakoshi. These data also confirm the continuous process of trophoblastic in-vasion throughout pregnancy.

We did not find any significant differences in ri, pi or sd-ratio in uncomplicatedpregnancies compared to pregnancies complicated by iugr. The same resultswere found by Hung and coworkers8: in their study of the spiral artery in week13-25 of gestation in pih or iugr complicated pregnancies, the pi did not differ sig-nificantly from uncomplicated pregnancies. We believe that the impairment of thetrophoblastic invasion in pregnancies complicated by iugr is probably not detectablein early pregnancy and therefore abnormal Doppler waveforms are most likelyfound later in pregnancy. The latter is confirmed by a study of Murakoshi10: fromgestational week 25 onwards he showed significantly higher ri’s in iugr compli-cated pregnancies compared to uncomplicated pregnancies and concludes thatDoppler flow velocity waveforms from gestational week 25 may identify thesepregnancies with growth restricted fetuses.

For pregnancies complicated by pe we again did not found any differences in


ri’s, pi’s or sd-ratio’s. When Matijevic and co-workers performed their measure-ments in the third trimester (gestational week 26 till 40), they did found higherimpedance to blood flow in spiral arteries in preeclamptic pregnancies comparedto uncomplicated pregnancies9. From their results they showed good diagnosticaccuracy for predicting adverse perinatal outcome. However, their results arebased on a prospective matched-pairs case controlled study, while our resultsare based upon prospective longitudinal data. In addition, they report that thetrophoblastic invasion is completed around 18-20 weeks of gestation and Dopplerultrasound measurements should be performed at least after this time. It is apossible explanation for our findings, since we performed our measurementsfrom week 11 till 24. Another explanation can be found in a study by Aardema et al.;he reports a difference between pe with or without good outcome and states thatonly pe with poor pregnancy outcome is associated with defective placentation,whereas pe with good outcome is not.16 In our pre eclamptic group there were nopoor pregnancy outcomes.

In pregnancies complicated by pih, we found significantly decreased imped-ance to flow (sd-ratio, pi and ri) preceding the clinical symptoms. We could notfind similar findings in the literature. Only Hung and coworkers mentioned therelationship of Doppler flow waveforms of spiral arteries and pih and they re-ported that no significant differences in these measurements were found.8 Ourresults are probably explained by the hyperdynamic disease model in which pre-pregnant women with preexisting hypertension and increased bmi have a hyper-dynamic circulation. This means they have a high cardiac output and com pensatoryvasodilatation in their spiral arteries to maintain near-normal blood pressure peripherally. This hyperdynamic circulation has been linked to endothelial activa-tion and subsequent hypertensive disorders during pregnancy.17-19 The maternalhemodynamics has been investigated by Bosio in a group of 400 primigravidas:women with pih showed a low resistance circulation and high cardiac output inthe preclinical phase and this state was maintained throughout pregnancy.20

Interestingly, in preeclamptic pregnancies a hemodynamic crossover after 24 weeksof gestation took place to low cardiac output and high resistance circulation. Un-fortunately, numbers of severely complicated pregnancies in our study weresmall and our measurements were performed till week 24 of gestation. Afterevaluation of our patients’ files, we found half of our pih complicated pregnan-cies suffering from pre-existing hypertension and significantly higher bmi, whichmay explain the significantly decreased impedance to flow we found and thusconfirming the hyperdynamic disease model.

This is the first study based on prospective longitudinal data, which we be-


lieves gives a more realistic view on spiral artery impedance changes comparedto cross-sectional or matched-pairs data. Our data was analyzed with multilevelanalysis. This new technique takes into account that the same women are re-peatedly measured and uses all the available data, irrespectively of the number ofrepeated measurements, which indicates that missing observations are allowed.Furthermore, multilevel analysis is capable of dealing with irregularly time inter-vals.

Matijevic and coworkers suggest that the measurements should include spiralarteries in both the central and peripheral region of the placenta since the ‘ringlike’ trophoblastic invasion is responsible for different impedance values be-tween these regions. They explain that the incomplete physiological change ofthe spiral arteries in the peripheral region will be of more significant differenceand therefore more suitable for detecting pregnancies complicated by pe.21 We hadchosen to perform our measurements in the central region of the placenta, firstto prevent any biases arising from measuring different parts of the placenta, sec-ondly because it’s difficult to distinguish peripheral from central placenta in earlypregnancy.

We are aware in this study that measuring two or three spiral arteries ineach examination is only a small number compared to the total amount (approx-imately 100) of spiral arteries. One could question the meaning of the informa-tion given by such a small number representing only a small percentage of thetotal uteroplacental perfusion. Also when obtaining Doppler flow velocity wave-forms of different spiral arteries we found fair intra-observer variability, whichshows the main difficulty of measuring spiral arteries; since there is a regional dif-ference in trophoblastic invasion and therefore a difference in modification of eachartery, measuring different arteries with Doppler can result in different sd-ratio,pi or ri values and therefore fair intraobserver variability. Increasing the amount ofmeasurements would probably give a better view of the ‘true’ impedance value, butwould increase the examination time for the patient to an unacceptable length.

Although significantly increased impedance to flow of spiral arteries in com-plicated pregnancies has been reported, we could not confirm this finding in earlypregnancy. Although our conclusion is based on a small number of cases, weconclude that early measurements of spiral artery flow velocity waveforms areprobably not sensitive enough to predict hypertensive disorders in pregnancy orpregnancies complicated by iugr. However, the longitudinal data based on ouruncomplicated pregnancies gives an interesting insight in spiral artery impedancechanges during normal pregnancy.


References

1 Pijnenborg R, Anthony J, Davey DA, Rees A,Tiltman A, Vercruysse L, van AA. 1991. Placentalbed spiral arteries in the hypertensive disordersof pregnancy. Br J Obstet Gynaecol 7: 648-655.

2 Brosens I.1964. A study of the spiral arteries of the decidua basalis in normotensive andhypertensive pregnancies. J Obstet Gynaecol Br Commonw 71: 222-230.

3 Meekins JW, Pijnenborg R, Hanssens M,McFadyen IR, van AA. 1994. A study of placentalbed spiral arteries and trophoblast invasion innormal and severe pre-eclamptic pregnancies.Br J Obstet Gynaecol 8: 669-674.

4 Brosens IA, Robertson WB, Dixon HG.1972.The role of the spiral arteries in thepathogenesis of preeclampsia. Obstet GynecolAnnu 1: 177-191.

5 Lyall F. 2005. Priming and remodelling of humanplacental bed spiral arteries during pregnancy– a review. Placenta 26 Suppl A: S31-S36.

6 Sheppard BL, Bonnar J. 1981. An ultrastructuralstudy of utero-placental spiral arteries inhypertensive and normotensive pregnancy andfetal growth retardation. Br J Obstet Gynaecol 7:695-705.

7 Jurkovic D, Jauniaux E, Kurjak A, Hustin J,Campbell S, Nicolaides KH. 1991. Transvaginalcolor Doppler assessment of the uteroplacentalcirculation in early pregnancy. Obstet Gynecol.77: 365-369.

8 Hung JH, Ng HT, Pan YP, Yang MJ, Shu LP. 1997.Color Doppler ultrasound, pregnancy-inducedhypertension and small-for-gestational-agefetuses. Int J Gynaecol Obstet 1: 3-11.

9 Matijevic R, Johnston T. 1999. In vivo assessmentof failed trophoblastic invasion of the spiralarteries in pre-eclampsia. Br J Obstet Gynaecol 1:78-82.

10 Murakoshi T, Sekizuka N, Takakuwa K,Yoshizawa H, Tanaka K. 1996. Uterine and spiralartery flow velocity waveforms in pregnancy-induced hypertension and/or intrauterinegrowth retardation. Ultrasound Obstet Gynecol 2:122-128.

11 Report of the National High Blood Pressure

Education Program Working Group on HighBlood Pressure in Pregnancy. 2000. Am J ObstetGynecol 183: S1-S22.

12 Twisk J. 2006. Applied Multilevel Analysis. A Practical Guide. Cambridge University Press:Cambridge.

13 Pijnenborg R, Bland JM, Robertson WB, Brosens I. 1983. Uteroplacental arterial changesrelated to interstitial trophoblast migration inearly human pregnancy. Placenta 4: 397-413.

14 Pijnenborg R, Vercruysse L, Hanssens M. 2006.The uterine spiral arteries in human pregnancy:Facts and controversies. Placenta 27: 939-958.

15 Hung JH, Ng HT, Pan YP, Yang MJ, Shu LP. 1997.Color Doppler ultrasound of spiral arteries innormal second-trimester pregnancies. ZhonghuaYi Xue Za Zhi (Taipei) 5: 289-294.

16 Aardema MW, Saro MC, Lander M, De WolfBT, Oosterhof H, Aarnoudse JG. 2004. Secondtrimester Doppler ultrasound screening of theuterine arteries differentiates betweensubsequent normal and poor outcomes ofhypertensive pregnancy: two differentpathophysiological entities? Clin Sci (Lond) 4:377-382.

17 Carr DB, McDonald GB, Brateng D, Desai M,Thach CT, Easterling TR. 2001. The relationshipbetween hemodynamics and inflammatoryactivation in women at risk for preeclampsia.Obstet Gynecol 6: 1109-1116.

18 Easterling TR. 1992. The maternalhemodynamics of preeclampsia. Clin ObstetGynecol 2: 375-386.

19 Easterling TR, Benedetti TJ, Schmucker BC,Carlson K, Millard SP. 1991. Maternalhemodynamics and aortic diameter in normaland hypertensive pregnancies. Obstet Gynecol 6:1073-1077.

20 Bosio PM, McKenna PJ, Conroy R, O‘Herlihy C.1999. Maternal central hemodynamics inhypertensive disorders of pregnancy. ObstetGynecol 6: 978-984.

21 Matijevic R, Meekins JW, Walkinshaw SA,Neilson JP, McFadyen IR. 1995. Spiral arteryblood flow in the central and peripheral areas of the placental bed in the second trimester.Obstet Gynecol 2: 289-292.


adam12s and pp13 as first trimesterscreening markers for adverse pregnancy outcome

Koen L DeurlooIngeborg H LinskensDirk J KuikAnnemieke C HeijboerMarinus A BlankensteinJohn MG van Vugt

Submitted

chapter 6

Abstract

objective:

Assess screening performance of first trimester serum measurements of A- Disintegrin-And-Metalloprotease 12-S (adam12s) and Placental Protein 13 (pp13) for preeclampsia (pe), gestational hypertension (gh) and small-for-gestational-age (sga) fetuses.

study design:

In this retrospective case-control study cases were matched for gestational andmaternal age at sampling. Results were expressed as multiples of the median(mom) and compared using Mann-Whitney test. Screening performance was assessed by Receiver-Operator-Characteristics (roc) curves.

results:

Seventeen cases of pe, 30 cases of gh and 8 cases of sga were matched with 165 controls. roc-analysis yielded area-under-the-curve (auc) for adam12s and pp13 of 0.63 and 0.59 for pe, 0.68 and 0.57 for gh and 0.59 and 0.62 for sga,respectively. Combined adam12 and pp13 did not improve auc. When specificityset at 80%, corresponding sensitivity of adam12s was 52% for gh.

conclusion:

Combined adam12s and pp13 measurements do not predict adverse pregnancyoutcome, but decreased first trimester adam12s levels are associated with gh.

Introduction

Preeclampsia, gestational hypertension and intrauterine growth restriction aremajor causes of maternal and fetal mortality and morbidity.¹ Although the pathogenesis is only partly understood, it is assumed that impaired trophoblastic invasionand subsequent placental insufficiency play an important role.2,3

Early monitoring of trophoblastic invasion is generally believed to enable usin identifying women at high risk for developing pregnancy complications that re-quire more intensive surveillance during pregnancy. In addition, early identificationof high risk pregnancies can be useful when designing clinical trials regarding earlytreatment or intervention to prevent fetal or maternal complications.

Over the last decade, various authors examined the potential of first trimestermaternal marker screening of impaired placental development to identify highrisk pregnancies, such as spiral artery Doppler measurements, uterine arteryDoppler measurements or maternal serum markers, such as soluble Flt, PlacentaGrowth Factor (PlGF), Pregnancy Associated Plasma Protein-A (papp-a) or freeBeta-Human Chorionic Gonadotrophin (free ß-hCG).4-9 However, large and evencontroversial differences in screening performance for detecting complicatedpregnancies are reported.

A Disintegrin And Metalloprotease 12-s (adam12s) is a placenta–derived gly-coprotein produced by the trophoblast, which is involved in growth and differen-tiation.10 Literature has reported low serum levels of adam12s precedingpregnancies complicated by preeclampsia or intrauterine growth restriction.9,11-14

Although one study showed no association between adam12s and preeclampsiaor fetal growth restriction.15

Placental protein 13 (pp13) is a small dimer protein involved in implantationand spiral artery modification.16 Previous studies have show a relationship of lowserum pp13 levels and subsequent development of preeclampsia.17,18 Some authorshave used additional markers to increase screening performance, such as uterineartery Doppler or uterine artery Doppler and serum levels of papp-a.19,20 How-ever, one report did not show a significant relationship between low levels ofpp13 and fetal growth restriction.21

To our knowledge, there are no reports published regarding the combinedpredictive value of adam12s and pp13. Therefore we performed a retrospectivematched case-control study to investigate the potential value of combined first-trimester maternal serum adam12s and pp13 measurements for prediction of pre - eclampsia (pe), gestational hypertension (gh) or small-for-gestational-age (sga)fetuses.

69CHAPTE R 6 ADAM12S AND PP13 AS FIRST TRIMESTER SCREENING MARKERS FOR ADVERSE PREGNANCY OUTCOME

Methods

A retrospective matched case-control study was performed at the VU UniversityMedical Center, Amsterdam. Maternal serum samples were collected between9+0 till 13+6 weeks of gestation as part of the routine first-trimester screeningprogram for Down Syndrome in 2004-2007 and unused material was stored at -20°C for research purposes.

Long time stability of pp13 and adam12s at -20°C has been described pre -vi-ously.18,20,21 Before starting our case-control study, we evaluated an in house protocol and assessed the stability of pp13 and adam12s. We found stable con-centrations for the entire study period. For adam12s the results of the stabilityprotocol are published elsewhere.22

Gestational age at sample date was based on both dating scan and first dayof last menstrual period. When a difference of less than 7 days existed, gesta-tional age was based upon the first day of the last period. If a difference of atleast 7 days existed, a second dating scan was performed 2 weeks later. Gesta-tional age was then averaged from both dating scans. Pregnancy course and out-come were obtained by filled in questionnaires and delivery room records as wellas examination of the medical records and scanning of medical databases. Allparticipating women in this study gave informed consent. The study was approvedby our local Medical Ethic Committee.

According to the criteria of the International Society for the Study of Hyper-tension in Pregnancy cases of gestational hypertension were identified. Gesta-tional hypertension was defined as two recordings of diastolic blood pressureabove 90 mmHg at least 4 hours apart in a previously normotensive woman. Thesame criteria were used to identify cases of preeclampsia, defined as gestationalhypertension combined with proteinuria exceeding 300 mg/24 h or two readingsof at least 2+ by dipstick on urinalysis after 20 weeks of gestation.23

sga was de-fined as birth weight below the tenth percentile.24 Each case was matched forgestational age at sampling and maternal age with three control cases. In addi-tion, medical records were studied for data concerning known confounders suchas smoking status, ethnicity, primigravida and maternal weight. Pregnancy out-come data was evaluated for peak diastolic pressure, gestational age at delivery,fetal gender and neonatal weight at birth. We excluded multiple pregnancies,serum samples with unknown gestational age and fetal abnormalities. Controlcases delivered at term a healthy baby.

Serum adam12s was measured blinded for clinical outcome, using a semi- automatically performed time-resolved immunofluorometic assay (Autodelfia,


PerkinElmer, Turku, Finland). Interassay cv for the adam12s assay was below 5%at all levels. Serum pp13 was measured blinded for clinical outcome, using a solidphase enzyme linked immunosorbent assay (elisa), (Autodelfia, PerkinElmer,Turku, Finland). Interassay cv for the pp13 assay was below 5% at all levels.

Patient characteristics of both cases and controls are presented as medians(range) and percentages (%) and tested for significance with Mann Whitney Uand chi-square tests. adam12s and pp13 results were expressed in multiples ofthe median (mom). Differences in adam12s and pp13’s mom’s between cases andcontrols were evaluated for significance with non-parametric Mann-Whitneytests on the non-transformed mom’s, since log-transforming the adam12s andpp13 mom’s did not improve the fitting curve and were not normally distributed.

Receiver-operator-characteristics (roc) curves were used to asses screen-ing performance for adam12s, pp13 and both combined. Values of p < 0.05 wereconsidered significant.

Results

We identified 17 cases of pe and 30 cases of gh and 8 cases of sga fetuses andmatched 165 control cases. Patient and pregnancy outcome characteristics areshown in table 1 and 2, and are consistent with the matching criteria described inthe Methods section. Because of the matching procedure no significant differ-ences in gestational age and maternal age at sample collection were found be-tween controls and cases. In addition, no differences between cases and controlsfor smoking, ethnicity, primigravida and maternal weight were found. Logistic regression analysis demonstrated no significant confounding from maternal age,ethnicity, smoking, primigravida, fetal gender and neonatal weight on adam12sand pp13 levels.

Outcome characteristics are described in Table 2. As expected, we foundsignificant increased peak diastolic pressure in pregnancies complicated by pe andgh compared to the controls. Median neonatal weight at birth in pregnanciescomplicated by sga-fetuses was significantly reduced compared to controls. Inaddition, we found significantly decreased gestational age at delivery in pregnan-cies complicated by pe or sga-fetuses.

In controls, median adam12s concentration was 405 nG/L (range 101 – 923

nG/L) compared to 324 nG/L (range 85 – 761 nG/L) in cases. In controls, medianpp13 concentration was 57.7 pG/L (range 15.6 – 147 pG/L) compared to 54.6 pG/L(range 22.2 – 110 pG/L) in cases. adam12s and pp13 are independent analytes, log


regression analysis showed very weak correlation (maximum r = 0.2) betweenboth markers in the four outcome groups.

Table 3 shows median mom values for adam12s and pp13 for controls and cases. adam12s mom-values are significantly reduced in gh cases.

Table 4 shows roc curve analysis for adam12s, pp13 and combined adam12sand pp13. Results are expressed by area under the curve (auc). Sensitivity wasassessed when specificity was set at 80%. Figure 1 shows the roc curves for com-bined adam12s and pp13 measurements for prediction of pe, gh and sga fetuses,as well as the roc curve for adam12s measurement for prediction of gh.

table 1. Maternal characteristics in the four outcome groups.

Controls CasesCharacteristic pe gh sga p-value

Number 165 17 30 8 -

Median ga at sampling in days 82.1 80.1 80 84.9 0.55(range) (63-97) (64-94) (63-96) (65-95)

Median maternal age in years at withdrawal 34.4 35.7 34.7 34.7 0.91(range) (19.9-41.2) (30.3-41.6) (27.5-43.4) (30.4-37.8)

Median maternal weight in kg at sampling withdrawal 66.1 69.8 72.3 64 0.16(range) (50-110) (58-77) (56-110) (58-78)

Ethnicity (% Caucasian) 95.6 93.8 100 100 0.65

Primigravida (%) 59.5 53.8 68.0 100 0.19

Non-smoking (%) 95.2 90.9 100 85,7 0.47

pe: preeclampsia; gh: gestational hypertension; sga: small for gestational age; ga: gestational age. * p-value < 0.05 by post hoc testing.


table 2. Pregnancy outcome characteristics in the four outcome groups.

Controls CasesCharacteristic pe gh sga p-value

Number of cases 165 17 30 8 -

Median peak diastolic tension in mm Hg 75.6 101.9* 94.3* 89.5 <0.05(range) (60-90) (90-120) (90-110) (80-115)

Median ga at delivery in days 288 266* 275 261* <0.05(range) (236-297) (201-292) (231-296) (190-290)

Fetal gender (% female) 47.8 14.3 56.3 28.7 <0.07

Median neonatal weight at birth in gram 3546 3015 3383 2196* <0.05(range) (1865-4710) (1085-4000) (1960-4265) (540-2820)

pe: preeclampsia; gh: gestational hypertension; sga: small for gestational age; ga: gestational age. * p-value < 0.05 by post hoc testing.


table 3. Median mom values for adam12s and pp13 in controls and cases.

Controls CasesOutcome Measure pe gh sga

Number 165 17 30 8

adam12s in mom’s 1.00 0.90 0.77* 0.88(range) (0.25 – 2.28) (0.25-2.26) (0.45-1.20) (0.48-1.85)

pp13 in mom’s 1.00 0.77** 0.95 0.89(range) (0.27-2.52) (0.45-1.60) (0.35-1.90) (0.41-1.54)

pe: preeclampsia; gh: gestational hypertension; sga: small for gestational age; mom: Multiple of Median; adam12s: A Disintegrin And Metalloprotease 12-S; pp13: Placental Protein 13. * P-value < 0.05; ** P-value = 0.07


figure 1

Receiver–operating characteristics curve for the prediction of preeclampsia (pe), gestationalhypertension (gh) and small-for-gestational-age (sga) fetuses using a combination of medianmom value of combined adam12s and pp13 measurements (area under the curve: 0.63 for pe,0.66 for gh and 0.60 for sga) and adam12s (area under the curve: 0.68 for gh).

table 4. Overview of Receiver-operating Characteristics curve analysis ofadam12s, pp13 and combined adam12s & pp13 for prediction of pe, gh and sga

fetuses expressed in auc. Sensitivity was assessed when specificity was set at80%.

Case Serum Marker Analyte treshold auc p-value SensitivityConcentration (mom’s)

pe adam12s 1.09 0.63 0.08 41%pp 13 1.35 0.57 0.21 24%adam12s & pp13 1.01 & 1.25 0.63 0.08 31%

gh adam12s 0.99 0.68 <0.01 52%pp 13 1.23 0.57 0.22 32%adam12s & pp13 0.97 & 1.11 0.66 0.06 48%

sga adam12s 1.35 0.59 0.41 26%pp 13 1.14 0.62 0.26 38%adam12s & pp13 0.95 & 1.02 0.60 0.34 31%

auc: Area under the Curve; pe: preeclampsia; adam12s: A Disintegrin And Metalloprotease 12-S; pp13: Placental Protein 13; gh: gestational hypertension; sga: small for gestational age.

Comment

This is the first study to our knowledge that assesses the predictive values ofcombined adam12s and pp13 measurements for pe and gh complicated pregnan-cies as well as pregnancies complicated by sga-fetuses. In this retrospectivematched case-control study, the main finding is a significantly decreased adam12smom in pregnancies complicated by gh. Detection rate is 52% when specificity isset at 80%, showing a fair association between gh and adam12s.

In addition, in pe or sga complicated pregnancies adam12s median mom’swere decreased, although not significantly. pp13 median mom’s were also de-creased in pe, gh and sga complicated pregnancies, although this neither was sig-nificant.

The strength of the study is that all samples were conducted from the samesample base of the national screening program for Down syndrome and there-fore included women are likely to be comparable between the controls and cas-


es. Controls and cases were matched for relevant confounding factors such asmaternal age and gestational age at sampling.

Our study is limited by the small number of cases; it would be most interest-ing in creating subgroups of early versus late onset preeclampsia and intra uterinegrowth restricted fetuses with fetal weight below 5th instead of 10th percentile.However, no sufficient data was present.

It is remarkable that a combination of adam12s and pp13 hardly improvespredictive values for pe and sga compared to single adam12s and pp13 measure-ments, since several studies have reported their individual relationship betweenadam12s or pp13 with pe and fetal growth restriction, while only two studiesshowed no relationship between pp13 and pe or adam12s with pe and fetal growthrestriction.15,21

For pp13 Chafetz et al. found good predictive values for pe with sensitivity of91% and specificity of 90% in a case-control study among 425 pregnant women.17

These values were also stated by Gonen et al. in a prospective study, in whichsensitivity was 80% with a fpr of 20%.18 Differences with our findings can be ex-plained by the difference in gestational age (ga) at sampling: Chafetz et al. per-formed sampling at a median ga of 10+6 weeks and Gonen et al. performedsampling in week 6+0 till 10+0 of gestation. It has been reported that early firsttrimester pp13 measurements might have better predictive values due to its rolein early placental development.25 The most recent study by Khalil et al. showedgood predictive values of sensitivity of approximately 90% with a false positiverate (fpr) of 20%.26 They included only women with a high a priori risk of pe, suchas chronic hypertension, pregestational diabetes or renal diseases which can ex-plain the differences with our study group which consisted mostly of women witha low a priori risk.

adam12s has been previously described as a potential predictive marker foradverse pregnancy outcome: Cowans et al. describes in a case-control studyamong 1705 cases versus 414 control cases significant lower levels of adam12s insmall-forgestational-age neonates (birthweight below 10th percentile) and sig -nificant elevated levels in large for gestational age neonates (birthweight above90

th percentile).11 The significant relationship between decreased adam12s levelsand pe has firstly been reported by Laigaard et al. in a case control study of 324

controls versus 160 cases of pe.12 We cannot fully explain the differences with ourresults, maybe the criteria for matching were more stringent in his study and thesmall number of cases in our study may have influenced our results. Moreover,Laigaard reports the use of a different assay to measure adam12s; the use of dif-ferent antibodies to detect adam12s might explain the differences in results found


between these two studies. This was also stated in a recent study by Wortelboeret al.27

Combined screening performance of adam12 and pp13 has not been describedin literature before. Other studies have shown combinations of pp13 with firsttrimester or second trimester uterine artery Doppler velocimetry and papp-a.19,20

Especially early onset preeclampsia is found to be highly detectable with thesecombinations. adam12s and uterine artery Doppler showed a sensitivity of 66%for pe. Pihl et al. combined adam12s with free ß-hCG and papp-a for prediction ofsmall-for-gestational-age neonates, the authors stated that this combination isreaching clinical relevance with a detection rate of 39% with a false positive rateof 10%.14 Poon et al. did not find improvement in predictive value of adversepregnancy outcome such as pe, gh or preterm delivery when adding adam12smeasurements to papp-a, uterine artery Doppler velocimetry and high a prioririsk based on obstetric history.15

In conclusion, we believe that decreased first trimester levels of adam12s maybe useful in early prediction of gh. Decreased levels of pp13 were not significantlycorrelated with adverse pregnancy outcome such as pe, gh or sga. Combiningadam12s and pp13 does not appear to improve screening performance.

Further studies should evaluate the predictive role of adam12s in complicatedpregnancies, possible in combination with other promising first trimester markers,such as papp-a or PlGF or second trimester uterine artery Doppler measure-ments. One can argue to select pregnancies with a high a priori risk to improvescreening performance.


References

1 Lyall F. The human placental bed revisited.Placenta 2002;23:555-62.

2 Meekins JW, Pijnenborg R, Hanssens M,McFadyen IR, van Asshe A. A study of placentalbed spiral arteries and trophoblast invasion innormal and severe pre-eclamptic pregnancies.bjog 1994;101:669-74.

3 Pijnenborg R, Anthony J, Davey DA et al.Placental bed spiral arteries in the hypertensivedisorders of pregnancy. bjog 1991;98:648-55.

4 Melchiorre K, Leslie K, Prefumo F, Bhide A,Thilaganathan B. First-trimester uterine arteryDoppler indices in the prediction of small-for-gestational age pregnancy and intrauterinegrowth restriction. Ultrasound Obstet Gynecol2009;33:524-9.

5 Deurloo KL, Spreeuwenberg MD, Bolte AC,Van Vugt JM. Color Doppler ultrasound of spiralartery blood flow for prediction of hypertensivedisorders and intra uterine growth restriction: a longitudinal study. Prenat Diagn 2007;27:1011-

6 De Vivo A, Baviera G, Giordano D, Todarello G,Corrado F, D’anna R. Endoglin, PlGF and sFlt-1as markers for predicting pre-eclampsia. ActaObstet Gynecol Scand 2008;87:837-42.

7 Smith GC, Crossley JA, Aitken DA et al.Circulating angiogenic factors in early pregnancyand the risk of preeclampsia, intrauterinegrowth restriction, spontaneous preterm birth, and stillbirth. Obstet Gynecol2007;109:1316-24.

8 Poon LC, Maiz N, Valencia C, Plasencia W,Nicolaides KH. First-trimester maternal serumpregnancy-associated plasma protein-A andpre-eclampsia. Ultrasound Obstet Gynecol2009;33:23-33.

9 Spencer K, Cowans NJ, Nicolaides KH. Lowlevels of maternal serum papp-a in the firsttrimester and the risk of pre-eclampsia. PrenatDiagn 2008;28:7-10.

10 Gilpin BJ, Loechel F, Mattei MG, Engvall E,Albrechtsen R, Wewer UM. A novel, secretedform of human adam12 (meltrin alpha)provokes myogenesis in vivo. J Biol Chem1998;273:157-66.

11 Cowans NJ, Spencer K. First-trimester adam12

and papp-a as markers for intrauterine fetalgrowth restriction through their roles in theinsulin-like growth factor system. Prenat Diagn2007;27:264-71.

12 Laigaard J, Sorensen T, Placing S et al. Reductionof the disintegrin and metalloprotease adam12

in preeclampsia. Obstet Gynecol 2005;106:144-9.13 Spencer K, Cowans NJ, Stamatopoulou A.

adam12s in maternal serum as a potential markerof pre-eclampsia. Prenat Diagn 2008;28:212-6.

14 Pihl K, Larsen T, Krebs L, Christiansen M. Firsttrimester maternal serum papp-a, beta-hCG andadam12 in prediction of small-for-gestational-age fetuses. Prenat Diagn 2008;28:1131-5.

15 Poon LC, Chelemen T, Granvillano O, PandevaI, Nicolaides KH. First-trimester maternal seruma disintegrin and metalloprotease 12 (adam12)and adverse pregnancy outcome. Obstet Gynecol2008;112:1082-90.

16 Visegrady B, Than NG, Kilar F, Sumegi B, ThanGN, Bohn H. Homology modelling andmolecular dynamics studies of human placentaltissue protein 13 (galectin-13). Protein Eng2001;14:875-80.

17 Chafetz I, Kuhnreich I, Sammar M, Tal Y, GiborY, Meiri H et al. First-trimester placental protein13 screening for preeclampsia and intrauterinegrowth restriction. Am J Obstet Gynecol2007;197:35-7.

18 Gonen R, Shahar R, Grimpel YI et al. Placentalprotein 13 as an early marker for pre-eclampsia:a prospective longitudinal study. bjog

2008;115:1465-72.19 Nicolaides KH, Bindra R, Turan OM et al.

A novel approach to first-trimester screeningfor early pre-eclampsia combining serum pp-13

and Doppler ultrasound. Ultrasound ObstetGynecol 2006;27:13-7.

20 Spencer K, Cowans NJ, Chefetz I, Tal J,Kuhnreich I, Meiri H. Second-trimester uterineartery Doppler pulsatility index and maternalserum pp13 as markers of pre-eclampsia. PrenatDiagn 2007;27:258-63.

21 Cowans NJ, Spencer K, Meiri H. First-trimestermaternal placental protein 13 levels inpregnancies resulting in adverse outcomes.Prenat Diagn 2008;28:121-5.


22 Linskens IH, Beertsen CM, van Vugt JM,Blankenstein MA, Heijboer AC. Processingeffects and storage conditions on A DisintegrinAnd Metalloprotease (adam12s), a maternalserum marker for adverse pregnancy outcome.Clin Chem Lab Med 2009; 47: 1579-81.

23 Brown MA, Lindheimer MD, de Swiet M, Van Assche A, Moutquin JM. The classificationand diagnosis of the hypertensive disorders of pregnancy: statement from the InternationalSociety for the Study of Hypertension inPregnancy. Hypertens Pregnancy 2001;20:ix-xiv.

24 Cetin I, Alvino G. Intrauterine growthrestriction: implications for placentalmetabolism and transport. A review. Placenta2009; 30 Suppl A: S77-82.

25 Huppertz B, Sammar M, Chefetz I, Neumaier-Wagner P, Bartz C, Meiri H. Longitudinaldetermination of serum placental protein 13during development of preeclampsia. FetalDiagn Ther 2008;24:230-6.

26 Khalil A, Cowans NJ, Spencer K, Goichman S,Meiri H, Harrington K. First trimester maternalserum placental protein 13 for the prediction ofpre-eclampsia in women with a priori high risk.Prenat Diagn 2009;29:781-9.

27 Wortelboer EJ, Linskens IH, Koster MP,Stoutenbeek P, Cuckle H, Blankenstein MA et al.adam12s as a first-trimester screening marker oftrisomy. Prenat Diagn 2009;29:866-9.


Decreased plasma levels of metastin in early pregnancy are associated with small for gestational age neonates

Eva ML SmetsKoen L DeurlooAtti T J J GoJohn MG van VugtMarinus A BlankensteinCees BM Oudejans

Prenatal Diagnosis2008; 28: 299-303

chapter 7

Abstract

objective:

To investigate whether pregnancies with small for gestational ge (sga)neonates, defined as customized birth weight below the 10th percentile,are associated with altered levels of metastin in maternal plasma in thefirst trimester.

methods:

Maternal blood was obtained between 8 and 14 weeks of pregnancy.Levels of metastin were measured in pregnancies with (n = 31) orwithout sga-neonates (n = 31), matched for gestational age atvenipuncture. Measurement of ß-hCG was included to study theinfluence of gestational age and placental volume on plasma levels of the measured markers.

results:

Metastin was significantly lower in sga-pregnancies compared to anequal number of matched uneventful pregnancies (metastin: 1376 ± 1317

pmol/L vs 2035 ± 1260 pmol/L, p = 0.035; mean ± standard deviation),ß-hCG levels were not different.

conclusion:

Metastin is significantly lower in maternal plasma in the first trimester, in pregnancies with sga-neonates. It might therefore be used in combination with other markers for risk estimation of growthimpairment in the first trimester.

Introduction

Neonates that are small for gestational age (sga) are at increased risk of perinatalmorbidity and mortality.1-4 Furthermore, these babies are at increased risk of de-veloping cardiovascular disease and type 2 diabetes later in adult life.2,5-7

Although pregnancies affected by sga-neonates and by intrauterine growthrestriction (iugr) are not synonymous, there is considerable overlap betweenboth conditions, certainly when customized birthweight centiles are used.3,4

iugr

suggests impairment of fetal growth in utero due to placental dysfunction; sga-neonates are defined as having birthweight below the 10th percentile. As such, asubstantial part of sga neonates are growth restricted,8 but sga will also includesome small but normal neonates.

Intrauterine growth restriction is not a disease entity, but a manifestation ofseveral fetal and maternal disorders. While iugr can result from endogenous fetal factors that affect fetal growth, such as chromosomal anomalies, congenitalmalformations and infections,9 the leading cause of iugr, with or without pre -eclampsia, is deficient transfer of nutrients and oxygen from mother to fetus,most commonly due to inadequate trophoblast invasion into the endometriumwith subsequent inadequate remodeling of maternal spiral arteries.10-16

The similarities between invasive trophoblasts and invasive tumor cells arestriking, but in contrast to metastasis, invasion of trophoblasts is both temporallyand spatially well controlled. The kiss-1 gene was shown to be a key factor incontrolling cancer cell dissemination with expression levels inversely correlatedwith invasion.17-20 Subsequently, the kiss-1 gene was found to encode a caboxy-terminally amidated peptide with 54 amino-acid residues (metastin).21-24 As in tu-mors, metastin appears to share a similar function in the placenta by inhibitingthe in vitro migration of the invasive extravillous trophoblast.21,24 In apparentcontrast to these in vitro findings, Bilban et al. found that expression of metastinin vivo was higher in the first compared to the third trimester trophoblast.24 Inother words, metastin levels peak in the period of pregnancy when the invasivepotential of the placenta is maximal and when regulation and limitation of tro-phoblastic invasion take place.25 In vivo, in contrast to in vitro, metastin levelsthus seem to parallel placental invasive capacity.

We therefore hypothesized that placentae with inadequate trophoblastic in-vasion, manifested clinically by sga-neonates, may have decreased plasma metastinlevels.

Measurements of ß-subunit of human chorionic gonadotrophin (ß-hCG) wasincluded to study the influence of gestational age, placental volume and secretoryactivity, all of which are known to be associated with ß-hCG.26-27

83CHAPTE R 7 DECREASED PLASMA LEVELS OF METASTIN IN EARLY PREGNANCY ARE ASSOCIATED WITH SMALL FOR GESTATIONAL AGE NEONATES

Methods and materials

PatientsThis protocol was approved by the Medical Ethics Committee of the VU UniversityMedical Center. All procedures were in accordance with the ethical standardsfor human experimentation established by the Declaration of Helsinki of 1975, re- vised in 1983.

We designed a one-to-one case-control study in which the first trimestermetastin levels were compared between women delivering sga neonates andwomen delivering appropriate for gestational age (aga) neonates.

Eight hundred and seventy pregnant women with gestational ages between 8and 14 weeks attending the outpatient clinic of Obstetrics of the VU UniversityMedical Center between June 2002 and July 2004 were enrolled in the study. Patients were recruited from two cohorts, being either at risk for preeclampsia oriugr (primiparity, history of preeclampsia or sga-neonates), or for Down syndromeas determined by first-trimester serum screening in combination with nuchaltranslucency and maternal age.28 Both cohorts yielded 242 and 628 patients re-spectively. By this, the patients recruited were representative for the high-risk pop -ulation attending an academic setting. All participants signed informed consent.Blood was drawn by venipuncture. Patients were asked to fill in a questionnaireon obstetrical history, genetic abnormalities, family history of obstetrical and fetalcomplications, life style, medication and disease, both inherited and acquired.The pregnancy outcome was checked after delivery by review of respective med-ical records. Exclusion criteria for both cases and controls were (1) incompletequestionnaire or missing medical records (n = 104); (2) multiple gestations (n = 8);(3) fetal chromosome anomalies (n = 4); (4) hypertensive disorders antedatingpregnancy (n = 34); pregnancy-induced hypertension (n = 43); preeclampsia orhemolysis elevated liver enzymes low platelets-syndrome (hellp) (n = 32); (5) preexisting disease not related to pregnancy (n = 5); and (6) smoking during preg-nancy (n = 2). Since hypertensive disorders are known to be associated withplacental insufficiency, patients affected by them were excluded from the populationunder investigation. Hypertensive disorders were defined as preexistent hyper-tension (blood pressure over 140/90 mmHg before the 20

th week of gestationmeasured on two separate occasions), pregnancy-induced hypertension (bloodpressure over 140/90 mmHg after the 20

th week of gestation measured on twoseparate occasions without preexisting hypertension), pre eclampsia (pregnancy-induced hypertension combined with proteinuria of 1+ or more on a dipstick ormore than 300 mg per day) and the hellp syndrome.


From the group of pregnancies resulting in delivery of aga neonates and fulfillingthe same exclusion criteria as the case group, we randomly selected controls ona one-to-one basis matching for gestational age at venipuncture within one day.Gestational age was confirmed by first-trimester ultrasound. Birth weight cen-tiles were corrected for physiological factors known to affect fetal growth: ma-ternal height, weight, ethnicity, parity, sex and gestational age at delivery sincethese customized birth weight centiles provide a better estimate of perinatalmorbidity and mortality than nonadjusted population centiles.3,4,29 Other physio-logical variables such as paternal height have, unless extreme, a relatively minoreffect. Maternal age appears to play no significant role once parity is controlledfor.30 If data for maternal height and weight were missing, the average of the studypopulation for that parameter was entered. Correction was done using specificsoftware based on an extensive Swedish database.30 Swedish and Dutch womenare comparable in maternal height and weight.31-35

sga-neonates were defined ashaving a customized birth weight below the 10th centile.

edta plasmaedta blood was stored at 4 °C and processed within 24 h after collection as de-scribed previously.36

SerumSerum was allowed to clot for a minimum of 1 h. After centrifugation it was storedat 4 °C and measured within 48 h.

MetastinMeasurements were carried out in edta plasma using the radio immunoassay kitfrom Phoenix Pharmaceuticals, Inc. (Blemont, ca, www.phoenixpeptide.com) as described in the product insert. One hundred microliters of plasma were in-cubated with 100 ml of anti-metastin-antibody for 16 – 24 hours at 4 °C. Then125I-kisspeptin-10 (amidated metastin 45-54) was added. Another 24 hours later,100 ml of goat-anti-rabbit-antiserum and rabbit serum were added and after incubation for 90 minutes, 500 ml buffer was added. Samples were decanted after centrifugation at 1700g for 20 minutes at 4 °C, and radioactivity was meas-ured.

Kisspeptin-10 (amidated metastin 45-54) was used to calibrate the assay. Re-activity of amidated metastin 1-54 was 100%. The calibration curve ranged from6 pg/ml to 1890 pg/ml. We limited the interpretation of results to the linear partof the calibration curve ranging from 38 to 770 pmol/L; only results within this


range were used in further calculations. Samples were diluted two- and eightfold.All measurements were performed in duplicate and averages were calculated.

All measurements were performed in the same run. Within-run coefficientof variation was 6.6% at 163 pg/ml and 7.3% at 347 pg/ml.

ß-subunit of human chorionic gonadotropin (ß-hCG)ß-hCG was measured in serum using the time-resolved fluoroimmunoassay kitmanufactured by Perkin Elmer (Waltham, ma). Analyses were carried out on theDelfia Xpress analyzer from the same manufacturer. Measurements were per-formed in different runs. Between-run coefficient of variation is below 5% forthe whole concentration range.

StatisticsPopulation characteristics are presented as mean ± standard deviation (sd). Onlymaternal height and weight were normally distributed and student’s t-tests wereused for comparison. For all other parameters, the nonparametric Mann-Whitneytest was used to compare the means of groups.

Correlation between two parameters was evaluated using Spearman’s cor-relation coefficient. Differences were considered statistically significant if p < 0.05.All statistical analyses were performed using Statistical Package for Social Sciences(spss) version 12.0.1. for Windows (spss, Inc., Chicago, il).

Results

Metastin was measured in maternal plasma between 8 and 14 weeks of pregnan-cy in 31 women with sga-pregnancies and in 31 matched patients with confirmeduneventful pregnancies. Patient characteristics are given in table 1. By definition,birth weight and customized centiles were significantly different in both groups.Mean maternal height was also statistically different between both groups, butthe difference was relatively small. Furthermore, the centiles, used to classify theneonates, were corrected for maternal height.

Complementary serum was used to analyze ß-hCG was available in 26 of the31 patients with affected pregnancies. Levels of ß-hCG were compared to 26

patients of the randomly selected from the matched control group. In pregnan-cies with sga-neonates, mean first-trimester levels of metastin in maternal plas-ma were found to be significantly lower compared to the group of matcheduneventful pregnancies. In contrast, no differences in mean ß-hCG levels were


noted between both groups (Table 2, figure 1). Metastin and ß-hCG were notcorrelated in sga-pregnancies (r = -0.224, p = 0.272), nor in affected pregnancies(r = -0.194, p = 0.324).

table 1. Population characteristics presented as mean ± sd.

Population characteristics Matched uneventful Pregnancies affectedpregnanciesa by sga

a

N 31 31Maternal age (years) 33.6 ± 2.8 34.0 ± 4.1Maternal weight (kg) 65 ± 5 65 ± 9Maternal height (cm) 173 ± 8 b 166 ± 6b

Gestational age at venipuncture (days) 85 ± 8 85 ± 9Gestational age at delivery (days) 279 ± 8 276 ± 11Birth weight (g) 3623 ± 334b 2665 ± 369b

Customized birth weight centile 55.5 ± 23.7b 3.7 ± 2.8b

a Pregnancies with small for gestational age (sga) neonates and uneventful pregnancies were matched for gestational age at venipuncture within a time frame of one day.b Significant difference between 31 pregnancies affected by sga and 31 matched uneventfulpregnancies (p < 0.05); sd: standard deviation.


table 2. Mean circulating concentrations ± sd in first trimester of pregnancieswith small for gestational age (sga) neonates and matched unaffected controlpregnancies (nonparametric Mann-Whitney test).

Parameter Controlª sgaª P

Metastin (pmol/ml) 2035 ± 1260 (n = 31) 1376 ± 1317 (n = 31) 0.035 b

ß-hCG (pg/ml) 62 ± 56 (n = 26) 61 ± 55 (n = 26) 0.539

ª Pregnancies affected by sga and uneventful pregnancies were matched for gestational age at venipuncture within a time frame of one day.b Significant difference between pregnancies affected by sga and matched uneventful pregnancies (p < 0.05); sd: standard deviation.


figure 1. (A) Box plot of concentrations of metastin (pmol/L) in first-trimester plasma of 31 pregnancies with neonates that were small for gestational (sga) and 31 uneventful pregnan-cies matched for gestational age at venipuncture. (B) Box plot of concentrations of ß-hCG(pg/mL) in first-trimester serum of 26 pregnancies with sga-neonates and 26 uneventful pregnancies matched for gestational age at venipuncture. Outliers (ο) are values between 1.5 InterQuartile Ranges (iqr’s) and 3 iqr’s from one end of the box. Extreme values (•) are values more than 3 iqr’s from one end of a box. * Significant difference between pregnanciesaffected by sga and matched uneventful pregnancies (p < 0.05).

Discussion

We have found that first-trimester plasma metastin levels are significantly lowerin gestations destined to deliver sga neonates compared with those destined todeliver aga neonates.

Although abnormal fetal growth does not become apparent before the secondhalf of pregnancy, our findings suggest that the initial disturbance of growth liesin the first trimester of pregnancy. Already at 12 weeks of gestation, placentaefrom pregnancies with sga-neonates are smaller, whether combined with pre -eclampsia or not.37

Our study design prevented confounding caused by maternal factors like hy-pertensive disorders of pregnancy, smoking or preexisting disease. As such, thesga-population we studied represented the form with placental origin. The processinvolved and the dysfunctional appears to be related to trophoblastic invasion, asin preeclampsia.10,14,16

Metastin has been found to be involved in trophoblast invasion, through in-teraction of the ligand metastin, which is abundantly expressed in the syncytiotro -phoblast of the placenta, with its receptor, expressed by invasive extravilloustrophoblast.21,23,24 Metastin inhibits invasion in vitro by almost 70% without affect-ing its proliferative capacity.24 In apparent contrast to these findings, the highestexpression of metastin in vivo has been found in the first trimester of pregnancywhen placental invasive capacity is maximal and regulation of trophoblastic inva-sion is most important.24,25 We therefore hypothesize that in vivo metastin is in-volved in fine-tuning of placental invasion, in which a high invasive capacity wouldbe counteracted by the inhibitory effect of high levels of metastin and a low ex-pression of metastin would signal low invasive capacity. In concordance with thishypothesis, we found significantly decreased levels of circulating metastin in mater-nal plasma in first trimester of pregnancies with sga-neonates, i.e. pregnancieswith reduced invasive capacity.

The decrease in metastin does not appear to reflect merely a reduced pla-cental volume since ß-hCG, which is known to be correlated with placental volumein the first trimester,26 is not affected. Nor it is likely that decreased concentra-tions of circulating metastin are simply due to reduced transcriptional activity invascular growth restriction27 as this would also have influenced ß-hCG levels.

In summary, we found lower circulating levels of metastin in first-trimesterpregnancies affected by sga-neonates while levels of ß-hCG were not different.Although metastin by itself does not have sufficient discriminative power to


identify pregnancies at risk, it might be used in combination with other markersto identify pregnancies at risk for future sga.

Further studies will be performed to investigate whether pregnancies com-plicated by early preeclampsia (before 34 weeks of gestation) with concomitantiugr, which are also characterized by insufficient placental invasion, are also as-sociated with decreased levels of circulating metastin.


References

1 Boehm G, Senger H, Braun W, Beyreiss K, Raiha NC. 1988. Metabolic differences between aga- and sga-infants of very lowbirthweight. I. Relationship to intrauterinegrowth retardation. Acta Paediatr Scand 77(1):19-23.

2 Bakketeig LS, Magnus P. 1992. Small-for-gestational (sga) definitions andassociated risks. Int J Technol Asses Health Care9(Suppl. 1): 139-146.

3 McCowan LM, Harding JE, Stewart AW. 2005.Customized birth-weight centiles predict sga

pregnancies with perinatal morbidity. bjog 112

(8): 1-26-1033.4 Figueras F, Figueras J, Meler E, et al. 2007.

Customised birthweight standards accuratelypredict perinatal morbidity. Arch Dis Child FetalNeonatal Ed 92(4):F277-F280.

5 Godfrey KM, Barker DJ. 2000. Fetal nutritionand adult disease. Am J Clin Nutr 71(5): (Suppl. 1344S – 1352S).

6 Barker DJ, Eriksson JG, Forsen T, Osmond C.Fetal origins of adult disease: strength of effects and biological basis. Int J Epidemiol2002;31:1235-9.

7 Barker DJ. The developmental origins of adultdisease. Eur J Epidemiol 2003;18:733-6.

8 Chatelain P. 2000. Children born with intra-uterine growth retardation (iugr) or small forgestational age (sga): long term growth andmetabolic consequences. Endocr Regul 34(1): 33 – 36.

9 Snijders RJ, Sherrod C, Gosden CM, Nicolaides KH. Fetal growth retardation:associated malformations and chromosomalabnormalities. Am J Obstet Gynecol1993;168:547-55.

10 Brosens I, Dixon HG, Robertson WB. Fetal growth retardation and the arteries of the placental bed. Br J Obstet Gynaecol1977;84:656-63.

11 Sheppard BL, Bonnar J. 1980. Ultrastructuralabnormalities of placental villi in placentae frompregnancies complicated by intrauterine fetalgrowth retardation: their relationship to

decidual spiral arterial lesions. Placenta 1(2): 145-156.

12 Gerretsen G, Huisjes HJ, Elema JD. 1981.Morphological changes of the spiral arteries in the placental bed in relation to pre-eclampsiaand fetal growth retardation. bjog 88(9): 876-881.

13 Sheppard BL, Bonnar J. 1981. An ultrastructuralstudy of uteroplacental spiral arteries inhypertensive and normotensive pregnancy andfetal growth retardation. bjog 88(7): 695-705.

14 Khong TY, De Wolf F, Robertson WB, BrosensI. Inadequate maternal vascular response toplacentation in pregnancies complicated by pre-eclampsia and by small-for-gestational ageinfants. Br J Obstet Gynaecol 1986;93:1049-59.

15 Sheppard BL, Bonnar J. Uteroplacental hemo -stasis in intrauterine fetal growth retardation.Semin Thromb Hemost 1999;25:443-6.

16 Matijevic R, Ward S, Bajoria R. 2002. Non-invasive method of evaluation of trophoblastinvasion of spiral arteries in MonochorionicTwins with discordant Birthweight. Placenta23(1): 93-99.

17 Lee JH, Miele ME, Hicks DJ, Phillips KK, TrentJM, Weissman BE, et al. KiSS-1, a novel humanmalignant melanoma metastasis-suppressorgene. J Natl Cancer Inst 1996;88:1731-7.

18 Lee J.H., Welch D.R. Suppression of metastasisin human breast carcinoma mda-mb-435 cellsafter transfection with the metastasis sup -pressor gene, kiss-1. Cancer Res 1997;57:2384-7.

19 Lee J.H., Welch D.R. Identification of highlyexpressed genes in metastasis-suppressedchromosome 6/human malignant melanomahybrid cells using subtractive hybridization anddifferential display. Int J Cancer 1997;71:1035-44.

20 Sanchez-Carbayo M, Capodieci P, Cordon-Cardo C. Tumor Suppressor Role of kiss-1 inBladder Cancer: Loss of kiss-1 Expression IsAssociated with Bladder Cancer Progression andClinical Outcome. Am J Pathol 2003;162:609-17.

21 Kotani M, Detheux M, Vandenbogaerde A,Communi D, Vanderwinden JM, Le Poul E, et al.The metastasis suppressor gene kiss-1 encodeskisspeptins, the natural ligands of the orphan G protein-coupled receptor gpr54. J Biol Chem2001;276:34631-6.


22 Muir AI, Chamberlain L, Elshourbagy NA,Michalovich D, Moore DJ, Calamari A, et al.axor12, a novel human G protein-coupledreceptor, activated by the peptide kiss-1. J Biol Chem 2001;276:28969-75.

23 Horikoshi Y, Matsumoto H, Takatsu Y, Ohtaki T,Kitada C, Usuki S, et al. Dramatic Elevation ofPlasma Metastin Concentrations in HumanPregnancy: Metastin as a Novel Placenta-Derived Hormone in Humans. J Clin EndocrinolMetab 2003;88:914-9.

24 Bilban M, Ghaffari-Tabrizi N, Hintermann E,Bauer S, Molzer S, Zoratti C, et al. Kisspeptin-10,a kiss-1/metastin-derived decapeptide, is aphysiological invasion inhibitor of primary humantrophoblasts. J Cell Sci 2004;117:1319-28.

25 Ursula H, Martin B, Martin K, Gernot D. 2007.Kisspeptins and the placenta: regulation oftrophoblast invasion. Rev Endocr Metab Disord8(1): 31-39.

26 Metzenbauer M, Hafner E, Hoefinger D, et al. 2001. Three-dimensional ultrasoundmeasure ment of the placental volume in earlypreg nancy: method and correlation withbiochemical parameters. Placenta 22(6): 602-605.

27 Vaiman D, Mondon F, Alexandra GD, MignotTM, Robert B, Rebourcet R, et al. Hypoxia-activated genes from early placenta are elevatedin Preeclampsia, but not in Intra-UterineGrowth Retardation. bmc Genomics 2005;6:111.

28 Schielen PC, van Leeuwen-Spruijt M,Belmouden I, Elvers LH, Jonker M, Loeber JG.Multi-centre first-trimester screening for Downsyndrome in the Netherlands in routine clinicalpractice. Prenat Diagn 2006;26:711-8.

29 Groom KM, Popp KK, North RA, McCowan LME. 2007. Small-for-gestational-ageinfants classified by customized or populationbirthweight centiles: impact of gestational age at delivery. Am J Obstet Gynecol 197(3):239.

30 Gardosi J and Francis A. Customised CentileCalculator-grow-Centile. [v5.1]. 2006. GestationNetwork. www.gestation.net.

31 Lissner L, Johansson SE, Qvist J, Rossner S,Wolk A. 2000. Social mapping of the obesityepidemic in Sweden. Int J Obes Relat MetabDisord 24(6): 801-805.

32 Visscher TL, Kromhout D, Seidell JC. 2002.

Long-term and recent time trends in theprevalence of obesity among Dutch men andwomen. Int J Obes Relat Metab Disord 26(9):1218-1224.

33 Onland-Moret NC, Peeter PHM, van Gils CH,et al. 2005. Age at menarche in relation to adultheight: the epic Study. Am J Epidemiol 162(7):623-632.

34 Statistics Netherlands-Centraal Bureau voorStatistiek. 2007. Dutch Population Statistics.Centraal Bureau voor Statistiek.

35 Statistics Sweden – Statistiska Centralbyran.2007. Swedish population statistics. StatistiskaCentralbyran.

36 Ng EKO, Leung TN, Tsui NBY, Lau TK, Panesar NS, Chiu RWK, et al. The Con cen -tration of Circulating Corticotropin-releasingHormone mRNA in Maternal Plasma Is Increasedin Preeclampsia. Clin Chem 2003;49:727-31.

37 Hafner E. Metzenbauer M, Hofinger D, et al.2003. Placental growth from the first to thesecond trimester of pregnancy in sga-foetusesand pre-eclamptic pregnancies compared tonormal foetuses. Placenta 24(4): 336-342.


General discussion, conclusionsand future prospects

chapter 8

Preeclampsia (pe) and intra uterine growth restriction (iugr) are major causes ofmaternal and fetal morbidity and mortality1,2. The incidence of pe varies between0.5 and 2.5% in western countries, but is much higher (approximately 10%) in de-veloping countries. Incidence of iugr obviously varies with definition at cut-offlevel of 2.3th, 5th or 10th centile3. The pathophysiology of pe and iugr is incom-pletely understood, but impairment of the endovascular trophoblastic invasionand spiral artery modification in early pregnancy as result of an exaggerated in-flammatory response and subsequent endothelial cell activation plays a crucial rolein the pathogenesis.

pe and iugr are associated with severe maternal complications, such as maternalkidney and liver failure, cerebral hemorrhage, thrombocytopenia and coronaryevents, as well as fetal complications like respiratory distress, intraventricular hem -orrhage, necrotizing enterocolitis, sepsis and hypoglycemia which can be due toboth (iatrogenic) premature delivery and intra uterine growth restriction. But pe

and iugr complicated pregnancies might also cause long term major and minorcomplications for both mother en newborn. Elevated maternal risk for cerebraland cardiovascular disease has been described by Bellamy et al. and McDonald etal.4, 5, and according to the Barker hypothesis, growth restricted fetuses mightcompensate for hostile intrauterine environment by endocrine metabolic repro-gramming which can lead to the metabolic syndrome later in life, presenting ashypertension, diabetes or as cardiac diseases6,7. In addition, various studies haveshown a relationship between growth-restricted fetuses and impaired cognitivefunctions at childhood8,9.

Therefore, for many decades research has focused on prediction and preventionof hypertensive disorders of pregnancy and its sequellae. Early identification of highrisk women is important because it allows us to explore secondary preventiveoptions.

Until now, the only known moderately effective preventive measure is lowdose aspirin in women with a history of preeclampsia and/or iugr. Although manycontradictory results have been reported, a large Cochrane systematic reviewdid show significant reduction of pe complicated pregnancies and small for gesta-tional age (sga) fetuses in a meta analysis of 32000 women10. Other interventionssuch as calcium or magnesium supplementation, fish oil, antioxidants (vitamin Cor E, selenium), or dietary interventions (nutritional advise or salt restriction) havebeen disappointing; results of lmwh are contradictory and the only known effectivetreatment for established preeclampsia and iugr is timed delivery.


As long as the cause of pe/iugr remains elusive it is commonly accepted that ac-curate identification and early diagnosis of women at risk, in combination withtimed intervention will eventually lead to an improvement of maternal and neona-tal outcome, both in the short and long term.

Our research was to assess the predictive value of various markers in the firsttrimester concerning the prediction of preeclampsia with or without iugr or iso-lated iugr in a high risk population of primigravida and women with a history ofpe or iugr.

As stated in the introduction, a multi marker screening seems the most promisingmethod for early detection of pe and isolated iugr. Numerous clinical, biophysical,and biochemical tests have been proposed for early prediction of preeclampsiaand iugr, but only few report predictive values which reaches clinical relevance,usually with moderate sensitivity and high specificity. There is urgent need forsafe, rapid and preferably non-invasive tests.

The major strength of our research is that we collected longitudinal data to studyDoppler measurements of the spiral and uterine artery. It allows us to give a real -istic view on predictive values for pe and isolated iugr by sonographic markers,and it accurately assesses the assumed relationship between spiral and uterineartery Doppler measurements. In addition, a novel combination of promisingserum markers is studied and the reproducibility of 3D Ultrasound measurementsis extensively examined, as well as its clinical use is reviewed.

Weaknesses of our research is the small number of women that actually developedpe and/or iugr in the longitudinal study concerning spiral and uterine arteries,which might be responsible for the moderate predictive values.

95CHAPTE R 8 GENERAL DISCUSSION, CONCLUSIONS AND FUTURE PROSPECTS

Conclusion of this Thesis

Three dimensional ultrasound volume measurements are highly reproducible and might be useful for early prediction of pe and isolated iugr

First and second trimester uterine artery Doppler measurements accurately reflect trophoblastic invasion

Spiral artery Doppler measurements in first and second trimester are poor predictors of adverse pregnancy outcome

Metastin might be usable for early prediction of iugr complicated pregnancies.

pp13 and adam12s are poor predictors of adverse pregnancy outcome, although adam12s seems to have a modest association with pih

Future prospects

UltrasoundIn our study the predictive value of spiral arteries Doppler measurement is assessed,as well as its assumed relationship with the uterine artery Doppler. In addition,the role of 3dus volume measurements of placental and fetal volume measure-ments is evaluated. Both sonographic measurements have limited predictive valueconcerning adverse pregnancy outcome and future research should focus on dif-ferent sonographic techniques for assessment of placental function and subsequentobstetric complications.

Due to the rapid development of ultrasound equipment since the start of ourheppi-trial (Early Prediction of Preeclampsia and iugr), new diagnostic tools, suchas 3D Power Doppler sonography and intravascular sonoelastography have beendeveloped which might be useful for early detection of preeclampsia or isolatediugr. These two promising tools are described below.

Three dimensional Power Doppler sonographyIn the late nineties a combination of 3dus and Doppler Flow was used to assessblood flow in various tissues and its potential has been promising. With this technique, called 3D power Doppler sonography, one or more 3dus acquired volumesamples are combined with power Doppler of selected tissue. In each acquiredvolume the ratio of color voxels (flow) to total voxels (flow and non-flow) is as-sessed and expressed as vascularization index (vi) or flow index (fi) in which zerorefers to no flow and 1 to complete flow. Although recently it has been publishedthat these measurements can considerably overestimate actual flow up to 44 times11,numerous publications have been evaluating its clinical relevance since the firstreport12.

It is used in many medical fields besides obstetrics and gynecology, such as in theevaluation of lymph node metastasis and survival in patient with gastric cancer13,for evaluation of post radiotherapy treatment in patients with prostate cancer14

or as a prognostic factor in patients with colon cancer.15

The use of 3D Power Doppler sonography has also been evaluated for gynecol-ogical indications, such as in discriminating benign or malign endometrium in post -menopausal women16 or fibroid evaluation after GnRH treatment17. It has alsofound its way in fertility diagnostic testing, such as evaluation of peri implantation


(sub)endometrial blood flow after artificial reproductive therapy (iui or ivf) andpregnancy rate.18, 19

Only a modest amount of obstetric studies have been published; Basgul et al.in 2007 used 3D Power Doppler sonography for the assessment of the cervicalcharacteristics in pregnant women, in which no differences in mass or vascularitywere found in nulliparous versus multiparous women20. In the same year, Rizzo et al.examined first trimester placental vascularization in pregnancies with aneuploidfetuses, and although he found significantly lower vascularization indices in aneu-ploid placentas, he suggests it is unlikely to improve the detection rate of fetuseswith trisomy 13, 18 or 21.21

Normal values of 3D Power Doppler sonography in uncomplicated pregnancieshas been the subject of several studies; in two recent publications it is reportedthat little to no differences in vi, fi and vascularization flow index (vfi) in relationto gestational age was found in uncomplicated pregnancies: Filho et al. studied283 patients from ga 26 -35 weeks and found a minor increment of fi with ad-vancing gestational age in which placental location has no influence on the threevascular indices (vi, fi, vfi) studied.22 The same results were found by Zalud et al.in 2007 in 199 uncomplicated pregnancies from ga 14 -25

23. Although in 295 normalpregnancies vi, fi and vfi showed constant distribution throughout gestation, aminor decrement in fi was found with advancing gestational age.24

In a small study of 13 iugr complicated pregnancies versus 208 uncomplicatedcontrols between ga 12 -40 weeks, a weak correlation between vi, fi and vfi andadvancing ga was seen, and good reproducibility and significant decreased vi, fi,vfi in iugr complicated pregnancies.25 Also, in 80 women, cigarette smoking wasrelated to decreased first trimester vascularization indices and decreased birthweight.26 In 2008 Guiot et al. found a significant relationship between vi and fi

and iugr-complicated pregnancies in a case control study among 15 controls and30 iugr cases.27

Very few studies have been published concerning the relationship of pre -eclampsia and 3D Power Doppler sonography: One Chinese case control studyof 36 healthy pregnant women and 44 preeclamptic and hypertensive pregnantwomen is reported which showed decreased placental perfusion by vi, fi and vfi

in severe preeclamptic women.28

Although most reports are case control studies, it is clear that the results arepromising. There is lack of large and longitudinal studies to assess screening per-


formance and this is needed to further elucidate the predictive value of 3D PowerDoppler sonography for early prediction of pe and/or iugr.

Sonographic elastographyDiscovered in 1990 by Ophir et al. it was at primarily used to identify breast lesions.29

Vascular elastography by ultrasound is a technique which assesses the strain inthe arterial wall and plaque by using an ultrasound image which is acquired at alow pressure (diastolic) and an elevated pressure (systolic or external compres-sion). The results are expressed as an elastogram which can reveal the presenceof an eccentric region with increased strain values and might therefore identifyplaques.

Experiments with this new technique during the research phase of this thesisdid not result in valuable results, probably due to the small diameter of the spiralarteries and the difficulty to use external compression. But based on previousexperience in in vitro studies30 and the expected advancement in ultrasoundequipment technology, it might be possible in the future to measure very smallarteries such as spiral arteries. Then it may become possible to differentiate be-tween well modified arteries with weak arterial walls (low-pressure) and incom-plete modified arteries with rigid arterial walls, which have increased resistanceindices due to shallow trophoblastic invasion. This might allow us to identify pe

or iugr complicated pregnancies in an early phase.

Serum markers

The predictive values of serum markers adam12s, pp13 and metastin is assessedin our research and although interesting results concerning the relationship with pih

and sga complicated pregnancies have been found, the clinical relevance seemslimited.

Since the start of our research in 2004, numerous maternal serum markers havebeen discovered and their relationship with adverse pregnancy outcome hasbeen assessed. Recent literature focuses on first trimester screening by variousmaternal serum markers: serum pregnancy-specific beta-1-glycoprotein (sp1), whichis associated with sga in a case control study.31 The proform of eosinophil majorbasic protein (proMBP), which is associated with pih, sga and pe

32 and glucoseregulated protein 78 (grp78)33, cystatine C34, matrix metalloproteinase-935, whichare associated with pe.


Multiple serum marker screening such as soluble endoglin, PlGF and sFlt36 or incombination with tgf-beta-137, or papp-a, beta-hCG and adam12

38 have been subjectof evaluation as well. For the prediction of iugr, a combination of PlGF, vegf,vegfr-1 and placental volume39, as well as first trimester ua pi and papp-a meas-urements40 have been studied. Predictive values differ considerably between thesestudies and although good screening performance has been reported, there isneed for prospective, large, multicenter, multi-ethnic studies to confirm thesepromising results41,42,36.

The use of serum markers is limited by the development of specific antibody assays, which is a timely and costly procedure. Since pe and isolated iugr seemgenetically determined diseases, understanding their genetics might help us finduseful predictive markers. Therefore, recent research also focuses on placentalgenetic products in maternal plasma, such as endovascular trophoblastic cells,cell-free fetal dna or rna.

Endovascular cells in maternal plasma are not informative due to their low con-centration in maternal blood (maximum of 1 cell in 1 ml maternal blood in the firsttrimester) and their year long survival in maternal blood. Therefore their clinicaluse is limited, especially in multiparous women.43

Cell-free dna or rna are shed in the circulation in large amounts and protectedfor months by micro particles, which makes them more feasible and accessiblefor screening purposes than endovascular trophoblastic cells.

Large amounts of cell-free rna and dna can be analysed and novel potentbiomarkers can be rapidly screened, including biomarkers not accessible by anti-body based assays.43,44 Genes can be selected with the help of the Sym Atlas, selecting only those with potential predictive value, which are placenta specificgenes with high expression levels compared to other maternal organs preferablyin the first trimester.

Useful applications of fetal dna or rna have been reported and include fetalgender determination45, fetal Rhesus typing46 and isolated gene disorders, such asHuntington’s disease47 and achondroplasia48. The use of fetal dna or rna for detection of aneuploid fetuses, especially trisomy 21 is more difficult, due to theadditional dna and rna production of the maternal chromosome 21

49, but it isstill the subject of current research.


The use of fetal dna or rna for identification of pe or iugr has been the subjectof few studies: in 1999 Lo et al. shows a five fold increase in maternal concentra-tion in preeclamptic pregnancies, probably due to increased apoptosis of the pla-cental cells.50 He suggests that cell free fetal dna could be a potent predictivemarker for preeclampsia. His results were confirmed by Zhong et al. in 2001.51

Van Dijk et al. reports that the placentally expressed stox1 gene is responsi-ble for the familial form of preeclampsia in Dutch women and combining thesemarkers with promising other potent biomarkers (from so called preeclampsiasusceptibility genes) would permit presymptomatic diagnosis of preeclampsia52,

53. The use of first trimester cell free fetal dna or rna measurements remains tobe explored.

Multi marker screening

It seems unlikely from the literature that one marker will be found that has enoughpredictive power, i.e. a combination of high specificity with high sensitivity, andtherefore recent studies focus on a combination of multiple predictive markers.This is a combination of sonographic markers, serum markers and/or Dopplerultrasound measurements for prediction of pe, such as first trimester pp13, papp-a

and second trimester ua pulsatility index by Spencer et al.41, sFlt and PlGF withsecond trimester ua pulsatility index42, papp-a, PlGF and ua pi

54 or first trimesterua pulsatility index and 3dus placenta volume measurement 55. These studies re-ports predictive values which reaches clinical relevance, although high specificityis generally accompanied by moderate to good sensitivity, which makes them notpractical in today’s clinical practice.

Conclusion

Preeclampsia and iugr are multifactorial diseases with unknown etiology. Impairedendovascular trophoblastic invasion plays a crucial role in the pathology, but theunderlying mechanisms are unclear. Although previously believed, there is not justone ‘X-factor’ responsible for these obstetric diseases and therefore it remains amajor challenge to identify women in early pregnancy that will develop pe or iugr.

When accurate prediction becomes available, the development and investigationof new preventive strategies is mandatory. However, as long as good preventive


measures are lacking, clinicians should focus on risk stratification of pregnant womanand adapt care accordingly. Low risk women should receive routine care, whilehigh risk women should receive adapted care according to their risk estimation.

Which combination of markers is best suitable for risk stratification still remainsunclear, but this research provides more insight in the selection of multi markersand the predictive values of current potent markers.


References

1 Sibai B, Dekker G, Kupferminc M. Pre-eclampsia. Lancet 2005; 9461: 785-799.

2 Khan KS, Wojdyla D, Say L, Gulmezoglu AM,Van Look PF. Who analysis of causes ofmaternal death: a systematic review. Lancet2006; 9516: 1066-1074.

3 Cetin I and Alvino G. Intrauterine growthrestriction: implications for placentalmetabolism and transport. A review. Placenta2009; S77-S82.

4 Bellamy L, Casas JP, Hingorani AD, Williams DJ.Pre-eclampsia and risk of cardiovascular diseaseand cancer in later life: systematic review andmeta-analysis. bmj 2007; 335: 974.

5 McDonald SD, Malinowski A, Zhou Q, Yusuf S,Devereaux PJ. Cardiovascular sequelae ofpreeclampsia/eclampsia: a systematic reviewand meta-analyses. Am Heart J 2008; 5: 918-930.

6 Barker DJ, Winter PD, Osmond C, Margetts B,Simmonds SJ. Weight in infancy and death fromischaemic heart disease. Lancet 1989; 8663: 577-580.

7 Kaijser M, Bonamy AK, Akre O, Cnattingius S,Granath F, Norman M, Ekbom A. Perinatal riskfactors for ischemic heart disease: disentanglingthe roles of birth weight and preterm birth.Circulation 2008; 3: 405-410.

8 Tideman E, Marsal K, Ley D. Cognitive functionin young adults following intrauterine growthrestriction with abnormal fetal aortic bloodflow. Ultrasound Obstet Gynecol 2007; 6: 614-618.

9 Puga B, Ferrandez LA, Garcia RR, Mayayo E,Labarta JI. Psychomotor and intellectualdevelopment of children born with intrauterinegrowth retardation (iugr). J Pediatr EndocrinolMetab 2004; 457-462.

10 Duley L, Henderson-Smart DJ, Knight M, King JF. Antiplatelet agents for preventing pre-eclampsia and its complications. CochraneDatabase Syst Rev 2004; 1: CD004659.

11 Schulten-Wijman MJ, Struijk PC, Brezinka C, De Jong N, Steegers EA. Evaluation of volumevascularization index and flow index: a phantomstudy. Ultrasound Obstet Gynecol 2008; 4: 560-564.

12 Pairleitner H, Steiner H, Hasenoehrl G,Staudach A. Three-dimensional power Dopplersonography: imaging and quantifying blood flowand vascularization. Ultrasound Obstet Gynecol1999; 2: 139-143.

13 Chen CN, Lin JJ, Lee H, Cheng YM, Chang KJ,Hsieh FJ, Lai HS, Chang CC, Lee PH.Association between color doppler vascularityindex, angiogenesis-related molecules, andclinical outcomes in gastric cancer. J Surg Oncol2009; 7: 402-408.

14 Strigari L, Marsella A, Canitano S, Gomellini S,Arcangeli S, Genovese E, Saracino B, PetrongariMG, Sentinelli S, Crecco M, Benassi M,Arcangeli G. Color Doppler quantitativemeasures to predict outcome of biopsies inprostate cancer. Med Phys 2008; 11: 4793-4799.

15 Chen CN, Cheng YM, Liang JT, Lee PH, Hsieh FJ, Yuan RH, Wang SM, Chang MF, Chang KJ. Color Doppler vascularity index can predict distant metastasis and survival incolon cancer patients. Cancer Res 2000; 11: 2892-2897.

16 Opolskiene G, Sladkevicius P, Jokubkiene L,Valentin L. Three-dimensional ultrasoundimaging for discrimination between benign and malignant endometrium in women withpostmenopausal bleeding and sonographicendometrial thickness of at least 4.5 mm.Ultrasound Obstet Gynecol 2010; 1: 94-102.

17 Chia CC, Huang SC, Chen SS, Kang JY, Lin JC,Lin YS, Huang KF, Lee HJ, Zheng CC.Ultrasonographic evaluation of the change inuterine fibroids induced by treatment with aGnRH analog. Taiwan J Obstet Gynecol 2006; 2: 124-128.

18 Kim A, Han JE, Yoon TK, Lyu SW, Seok HH,Won HJ. Relationship between endometrial and subendometrial blood flow measured bythree-dimensional power Doppler ultrasoundand pregnancy after intrauterine insemination.Fertil Steril 2009; (Epub ahead of print).

19 Ng EH, Chan CC, Tang OS, Yeung WS, Ho PC.Changes in endometrial and subendometrialblood flow in ivf. Reprod Biomed Online 2009; 2: 269-275.

20 Basgul A, Kavak ZN, Bakirci N, Gokaslan H.Three-dimensional ultrasound power Doppler


assessment of the cervix: comparison betweennulliparas and multiparas. J Perinat Med 2007; 1: 48-50.

21 Rizzo G, Capponi A, Cavicchioni O, Vendola M,Arduini D. Placental vascularization measured by three-dimensional power Doppler ultrasoundat 11 to 13 + 6 weeks’ gestation in normal andaneuploid fetuses. Ultrasound Obstet Gynecol2007; 3: 259-262.

22 Guimaraes Filho HA, Mattar R, Araujo JE, Da Costa LL, Mello Junior CF, Nardozza LM,Nowak PM, Moron AF. Reproducibility ofthree-dimensional power Doppler placentalvascular indices in pregnancies between 26 and35 weeks. Arch Gynecol Obstet 2010; (Epubahead of print)

23 Zalud I and Shaha S. Evaluation of the utero-placental circulation by three-dimensionalDoppler ultrasound in the second trimester ofnormal pregnancy. J Matern Fetal Neonatal Med2007; 4: 299-305.

24 de Paula CF, Ruano R, Campos JA, Zugaib M.Quantitative analysis of placental vasculature by three-dimensional power Doppler ultra -sonography in normal pregnancies from 12 to 40

weeks of gestation. Placenta 2009; 2: 142-148.25 Noguchi J, Hata K, Tanaka H, Hata T. Placental

vascular sonobiopsy using three-dimensionalpower Doppler ultrasound in normal andgrowth restricted fetuses. Placenta 2009; 5: 391-397.

26 Rizzo G, Capponi A, Pietrolucci ME, Arduini D.Effects of maternal cigarette smoking onplacental volume and vascularization measuredby 3-dimensional power Doppler ultra -sonography at 11+0 to 13+6 weeks of gestation.Am J Obstet Gynecol 2009; 4: 415-415.

27 Guiot C, Gaglioti P, Oberto M, Piccoli E, Rosato R, Todros T. Is three-dimensional powerDoppler ultrasound useful in the assessment ofplacental perfusion in normal and growth-restricted pregnancies? Ultrasound ObstetGynecol 2008; 2: 171-176.

28 Ma XQ, Wu QQ, Li P, Wang Q, Li M.Investigate of three-dimensional powerDoppler ultrasound used in placental perfusionassessment in pre-eclampsia pregnancies.Zhonghua Fu Chan Ke Za Zhi 2009; 3: 179-182.

29 Ophir J, Alam SK, Garra B, Kallel F, KonofagouE, Krouskop T, Varghese T. Elastography:ultrasonic estimation and imaging of the elasticproperties of tissues. Proc Inst Mech Eng H 1999;3: 203-233.

30 de Korte CL, Ignacio Cespedes EI, van der SteenAF, Lancee CT. Intravascular elasticity imagingusing ultrasound: feasibility studies in phantoms.Ultrasound Med Biol 1997; 5: 735-746.

31 Pihl K, Larsen T, Laursen I, Krebs L, ChristiansenM. First trimester maternal serum pregnancy-specific beta-1-glycoprotein (sp1) as a marker ofadverse pregnancy outcome. Prenat Diagn 2009;13: 1256-1261.

32 Pihl K, Larsen T, Rasmussen S, Krebs L,Christiansen M. The proform of eosinophilmajor basic protein: a new maternal serummarker for adverse pregnancy outcome. PrenatDiagn 2009; 11: 1013-1019.

33 Laverriere A, Landau R, Charvet I, Irion O,Bischof P, Morales M, Cohen M. grp78 as amarker of pre-eclampsia: an exploratory study.Mol Hum Reprod 2009; 9: 569-574.

34 Thilaganathan B, Ralph E, Papageorghiou AT,Melchiorre K, Sheldon J. Raised maternal serumcystatin C: an early pregnancy marker for pre -eclampsia. Reprod Sci 2009; 8: 788-793.

35 Poon LC, Nekrasova E, Anastassopoulos P,Livanos P, Nicolaides KH. First-trimestermaternal serum matrix metalloproteinase-9(mmp-9) and adverse pregnancy outcome.Prenat Diagn 2009; 6: 553-559.

36 De Vivo A, Baviera G, Giordano D, Todarello G,Corrado F, D’anna R. Endoglin, PlGF and sFlt-1as markers for predicting pre-eclampsia. ActaObstet Gynecol Scand 2008; 8: 837-842.

37 Lim JH, Kim SY, Park SY, Yang JH, Kim MY, RyuHM. Effective prediction of preeclampsia by acombined ratio of angiogenesis-related factors.Obstet Gynecol 2008; 6: 1403-1409.

38 Pihl K, Larsen T, Krebs L, Christiansen M. Firsttrimester maternal serum papp-a, beta-hCG andadam12 in prediction of small-for-gestational-age fetuses. Prenat Diagn 2008; 12: 1131-1135.

39 Semczuk-Sikora A, Krzyzanowski A, StachowiczN, Robak J, Kraczkowski J, Kwiatek M, Semczuk M. [Maternal serum concentration of angiogenic factors: pigf, vegf and vegfr-1


and placental volume in pregnanciescomplicated by intrauterine growth restriction].Ginekol Pol 2007; 10: 783-786.

40 Pilalis A, Souka AP, Antsaklis P, Daskalakis G,Papantoniou N, Mesogitis S, Antsaklis A.Screening for pre-eclampsia and fetal growthrestriction by uterine artery Doppler and papp-a

at 11-14 weeks’ gestation. Ultrasound ObstetGynecol 2007; 2: 135-140.

41 Spencer K, Cowans NJ, Chefetz I, Tal J, Meiri H.First-trimester maternal serum pp-13, papp-a

and second-trimester uterine artery Dopplerpulsatility index as markers of pre-eclampsia.Ultrasound Obstet Gynecol 2007; 2: 128-134.

42 Diab AE, El Behery MM, Ebrahiem MA, Shehata AE. Angiogenic factors for theprediction of pre-eclampsia in women withabnormal mid trimester uterine artery Dopplervelocimetry. Int J Gynaecol Obstet 2008; 2: 146-151.

43 Smets EM, Visser A, Go AT, van Vugt JM,Oudejans CB. Novel biomarkers inpreeclampsia. Clin Chim Acta 2006; 1-2: 22-32.

44 Oudejans CB, van Dijk M, Oosterkamp M,Lachmeijer A, Blankenstein MA. Genetics ofpreeclampsia: paradigm shifts. Hum Genet 2007;5: 607-612.

45 Stanghellini I, Bertorelli R, Capone L, Mazza V,Neri C, Percesepe A, Forabosco A.Quantitation of fetal dna in maternal serumduring the first trimester of pregnancy by theuse of a daz repetitive probe. Mol Hum Reprod2006; 9: 587-591.

46 Gautier E, Benachi A, Giovangrandi Y, Ernault P,Olivi M, Gaillon T, Costa JM. Fetal RhDgenotyping by maternal serum analysis: a two-year experience. Am J Obstet Gynecol 2005; 3: 666-669.

47 Gonzalez-Gonzalez MC, Trujillo MJ, RodriguezdA, Ramos C. Early Huntington disease prenataldiagnosis by maternal semiquantitativefluorescent-pcr. Neurology 2003; 7: 1214-1215.

48 Saito H, Sekizawa A, Morimoto T, Suzuki M,Yanaihara T. Prenatal dna diagnosis of a single-gene disorder from maternal plasma. Lancet2000; 9236: 1170-

49 Lo YM, Lau TK, Zhang J, Leung TN, Chang AM,Hjelm NM, Elmes RS, Bianchi DW. Increased

fetal dna concentrations in the plasma ofpregnant women carrying fetuses with trisomy21. Clin Chem 1999; 10: 1747-1751.

50 Lo YM, Leung TN, Tein MS, Sargent IL, Zhang J,Lau TK, Haines CJ, Redman CW. Quantitativeabnormalities of fetal dna in maternal serum inpreeclampsia. Clin Chem 1999; 2: 184-188.

51 Zhong XY, Laivuori H, Livingston JC, Ylikorkala O, Sibai BM, Holzgreve W, Hahn S.Elevation of both maternal and fetalextracellular circulating deoxyribonucleic acidconcentrations in the plasma of pregnantwomen with preeclampsia. Am J Obstet Gynecol2001; 3: 414-419.

52 van Dijk M, Mulders J, Poutsma A, Konst AA,Lachmeijer AM, Dekker GA, Blankenstein MA,Oudejans CB. Maternal segregation of theDutch preeclampsia locus at 10q22 with a newmember of the winged helix gene family. NatGenet 2005; 5: 514-519.

53 Oudejans CB and van Dijk M. Placental geneexpression and pre-eclampsia. Placenta 2008;S78-S82.

54 Poon LC, Kametas NA, Maiz N, Akolekar R,Nicolaides KH. First-trimester prediction ofhypertensive disorders in pregnancy.Hypertension 2009; 5: 812-818.

55 Rizzo G, Capponi A, Cavicchioni O, Vendola M,Arduini D. First trimester uterine Doppler andthree-dimensional ultrasound placental volumecalculation in predicting pre-eclampsia. Eur J Obstet Gynecol Reprod Biol 2008; 2: 147-151.


SummarySamenvatting

chapter 9

Summary

In the introduction (chapter 1) an overview of relevant predictive markers forpreeclampsia and intra uterine growth restriction is described. chapter 2 focuseson reproducibility of three-dimensional ultrasound measurements and in chapter 3the clinical use of sonographic fetal and placental volumes measurements is re-viewed. chapter 4 describes the relationship of spiral and uterine artery Dopplermeasurements and in chapter 5 the predictive value of spiral arteries Dopplermeasurements for adverse pregnancy outcome is assessed. In chapter 6 and 7maternal serum markers A Disintegrin And Metalloprotease 12-s (adam12s), pla-cental protein 13 (pp13) and metastin are evaluated for their predictive value forpreeclampsia (pe), pregnancy-induced hypertension (pih) and intra uterine growthrestriction (iugr). According to the criteria of the International Society for theStudy of Hypertension in Pregnancy 2001, cases of pregnancy-induced hyperten-sion (pih) were identified. Pregnancy-induced hypertension was defined as tworecordings of diastolic blood pressure above 90 mmHg at least 4 hours apart in apreviously normotensive woman. The same criteria were used to identify casesof preeclampsia, defined as pregnancy-induced hypertension combined with proteinuria exceeding 300 mg/24 h or two readings of at least 2+ by dipstick on uri-nalysis after 20 weeks of gestation.22 Intra uterine growth restriction was defined asbirth weight below the tenth percentile, customized by gestational age.

In chapter 2 we assessed the reproducibility of three-dimensional ultrasound(3dus) measurements of fetal and placental volumes in week 11 till 18 of gestation ina group of 34 uncomplicated pregnancies. Two operators independently acquiredfetal and placental volumes using 3dus. Each placental or fetal volume was ac-quired twice. Intra- and interobserver reproducibility was expressed in intra- andinterclass correlation coefficients (intra-cc and inter-cc). In addition, the effectof individual volume acquisition and caliper placement was evaluated.

Fetal and placental volume measurements were successful in 97% of all casesand the volumes were similar to previous reported literature. The intraobserverreproducibility was good for fetus (intra-cc: 0.99; 0.99) and placenta (intra-cc:0.99; 0.98). Also, interobserver reproducibility was good for fetus (inter-cc 0.98)and placenta (inter-cc 0.98). In addition, regardless of the operator who acquiredthe volumes, inter-cc remained good for both fetus (inter-cc: 0.99; 0.99) andplacenta (inter-cc: 0.97; 0.99).


Therefore, we concluded that intra- and interobserver variability of fetal and placental volume measurements are low and reproducibility is very good. In ad-dition, we have shown that individual volume acquisition and caliper placementhas no significant effect on the calculation of three-dimensional fetal and placen-tal volume measurements.

In chapter 3 the clinical use of 3dus measurements of placenta and fetal volumeis reviewed. Three-dimensional ultrasound fetal volume measurements are basedon head-and-trunk volumes, they are highly reproducible and seem very accurate inassessing fetal weight and monitor fetal growth. However, fetal volume measure-ments are limited by gestational age, since late second and third trimester fetusesusually exceed the size of the volume box. The clinical value of first tri mestermeasurements is yet to be determined, but significant decreased fetal volumeshave been reported in aneuploid fetuses.

Placental volumes measurements are also highly reproducible and easily ac-quired by 3dus. Like fetal volumes, these placental volume measurements arelimited by gestational age. Literature reports significantly smaller placental vol-umes in aneuploid fetuses, such as trisomy 13, 18 and 21, although its predictivevalue seems limited and addition of first trimester placental volume measure-ments to the current screening tests for aneuploid fetuses shows no improvementin detection rates. The relationship between second trimester placental volumeand pregnancies complicated by hypertensive disorders or sga-fetus seems morepromising, although its predictive values are limited by the large heterogeneity inplacental growth and size.

Further research for the assessment of predictive values of first trimester fetaland placental volume measurements is warranted.

chapter 4 reports the longitudinal relationship between Doppler flow velocitywave forms of the spiral artery (sa) and uterine artery (ua) in pregnant women.In this study, we analyzed 97 primigravidas with uncomplicated singleton pregnan -cies. In each pregnancy, sa and combined ua velocity waveforms were assessed usingtransabdominal ultrasound Color Doppler between gestational weeks 11 through14, 14 through 18 and 18 through 24. Each Doppler measurement was performedtwice. The combined ua was derived from the average pulsatility index (pi) of theleft and right ua. In addition, the presence of ua bilateral notching was reported.In total, 284 ua and 263 sa Doppler flow measurements were analyzed. Resultsshowed a continuous decrease of mean pi in sa and ua with increasing gestationalage and ua bilateral notching was reported in 35%, 9% and 3% of the cases be-

109CHAPTE R 9 SUMMARY

tween gestational weeks 11 through 14, 14 through 18 and 18 through 24 respec-tively. Intra observer variability for sa and combined ua was 0.54 and 0.90 re-spectively. Generalized Estimating Equations analysis showed significant correlation(r = 0.41) between sa and ua (p < 0.0001).

Our longitudinal data shows that the generally assumed relationship betweenuterine and spiral arteries Doppler measurements is confirmed and a significantcorrelation between these measurements is demonstrated; ua Doppler meas-urements in early pregnancy seem to accurately reflect peripheral resistance ofspiral arteries. Furthermore, this study shows that trophoblastic invasion seems acontinuous process in the first half of pregnancy in which first trimester bilateralnotching probably is a physiological phenomenon.

In chapter 5 we evaluated the longitudinal relationship between sa Dopplermeasurement and pregnancies complicated by pregnancy-induced hypertension(pih) or preeclampsia (pe), and intra uterine growth restricted (iugr) fetuses in108 pregnant women.

In all pregnancies sa blood flows were measured three times using ColorDoppler ultrasound. Measurement were performed between 11 and 14 week, be - tween 14 and 18 weeks and between 18 and 24 weeks of gestation, each measure-ment was performed twice. Spiral artery blood flows over time were analyzedwith multilevel modeling and reference ranges were constructed. Mann- Whitneytests were used to compare uncomplicated and complicated pregnancies. Weanalyzed 86 uncomplicated pregnancies and 21 complicated pregnancies (4 preg-nancies complicated by pe, 7 complicated by pih and 10 complicated by isolatediugr).

In the uncomplicated pregnancies, systolic/diastolic (sd-) ratios, resistanceindex (ri) as well as pulsatility index (pi) decreased progressively with advancinggestational age. Linear regression analysis showed that mean predicted sd-ratiodecreased from 1.75 at 11 weeks of gestation (P5 – P95: 1.32 – 2.17) to 1.48 at 24

weeks of gestation (P5 – P95: 1.05 – 1.90). The predicted ri en pi showed similardecrement: mean ri from 0.44 (P5 – P95: 0.28 – 0.59) at 11 weeks of gestation to0.34 (P5 – P95: 0.19 – 0.50) at 24 weeks of gestation, mean pi from 0.59 (P5 – P95:0.32 – 0.85) at 11 weeks of gestation to 0.40 (P5 – P95: 0.14 – 0.66) at 24 weeksof gestation. These results are consistent with previously reported literature. TheMann-Whitney tests showed no differences in pi, ri or sd-ratio in pregnancies withgrowth-restricted fetuses or complicated by preeclampsia. However, a significantlylower ri, pi and sd-ratio were found preclinically in pregnancies complicated bypih.


In conclusion, these results showed that sa Doppler measurements decreaseprogressively with advancing gestational age, but the reference ranges are wide.Therefore, sa Doppler measurements are probably not useful for early predic-tion of pe or iugr. Decreased Doppler indices preceding pih are likely caused bya hyperdynamic circulation in women with increased Body Mass Index and (un-masked) pre-existent hypertension.

In chapter 6, maternal serum markers adam12s and pp13 were evaluated fortheir first trimester predictive performance for adverse pregnancy outcome. Aretrospective case-control study of samples taken between 2004-2007 was conducted and 55 cases of pregnancies complicated by pe (n = 17), pih (n = 30)or isolated iugr fetuses (n = 8) were matched with 165 uncomplicated controlcases. The cases were matched for exact gestational age and maternal age withthree control cases. In addition, known confounders such as ethnicity, primi-gravida, smoking and maternal weight were analyzed. The serum concentrationof adam12s and pp13 were analyzed ‘blind’ to outcome and results were ex-pressed in multiples of the median (mom). There was not a normal distributionof adam12s and pp13 serum levels, therefore mom values were compared usingMann-Whitney U test. In addition, receiver-operator-characteristics (roc) curveswere used to asses screening performance.

Results showed significantly reduced median adam12s concentrations for con-trols versus all cases: 405 vs. 324 nG/L (mom 1.00 vs 0.80 (p < 0.05)). In pp13 nosignificant difference was found between controls and cases; 57.7 vs. 54.6 pG/L(mom 1.00 and 0.95). Multivariate analysis for possible dependency betweenadam12s and pp13 showed found very weak or no correlation between bothmarkers in the controls or in each group of cases.

Median mom levels for adam12s were 0.90, 0.77 and 0.88 for pe, pih and iugr

respectively; mom levels for pp13 were 0.77, 0.95 and 0.89 respectively. roc

analysis yielded areas under the curve (auc) for adam12s and pp13 of 0.63 and0.59 for pe, 0.68 and 0.57 for pih and 0.59 and 0.62 for iugr, respectively. Com-bined adam12 and pp13 did not improve auc. If specificity was set at 0.80, thecorresponding sensitivity of adam12s was 52% for pih.

This study demonstrated that adam12s was significantly associated withpregnancies complicated by pih, although predictive value was limited. De-creased levels of pp13 were not significantly correlated with adverse pregnancyoutcome. Combining adam12s and pp13 did not appear to improve screeningperformance.

111CHAPTE R 9 SUMMARY

In chapter 7 we assessed the association between first trimester plasma levelsof metastin and small for gestational age (sga) neonates, defined as customizedbirth weight below the 10th percentile. Maternal plasma was obtained betweenweek 8 till 14 weeks of gestation and levels of metastin were measured. Thirty-onecases of pregnancies with sga-neonates were matched with 31 pregnancies with-out sga-neonates for gestational age at venipuncture. Measurements of ß-hCGwere included to study the influence of gestational age and placental volume onplasma levels of the measured marker. Results showed significantly lower metastinlevels in sga-pregnancies compared to the uneventful pregnancies (metastin:1376 ± 1317 pmol/L vs 2035 ± 1260 pmol/L, p = 0.035), ß-hCG levels were notsignificantly different.

Therefore, we concluded that metastin might be useful in combination withother maternal serum markers for risk estimation of growth impairment in thefirst trimester.


Samenvatting

In de introductie (chapter 1) wordt een overzicht gegeven van relevante voor-spellende markers voor preëclampsie (pe) en intra-uteriene groeirestrictie (iugr).De focus van chapter 2 ligt op de reproduceerbaarheid van driedimensionaleechoscopische volumemetingen van foetus en placenta en in chapter 3 wordt deklinische relevantie van echoscopische foetale en placentaire volumemetingenbeschreven. De relatie tussen de spiraal arteriën en de uteriene arteriën is hetonderwerp van chapter 4 en in chapter 5 wordt de voorspellende waarde vanEcho Doppler metingen van de spiraal arteriën voor pe en iugr beoordeeld. Inchapter 6 and 7 worden de maternale serummarkers A Disintegrin And Metal-loprotease 12-s (adam12s), placental protein 13 (pp13) en metastin in relatie tot pe,pregnancy-induced hypertension (pih) en iugr geëvalueerd.

Volgens de criteria van the International Society for the Study of Hypertension inPregnancy (isshp 2001), worden cases met pe en pih geïdentificeerd: Pregnancy-induced hypertension is gedefinieerd als twee bloeddrukmetingen, waarbij de diastolische druk boven de 90 mm Hg ligt met ten minste vier uur interval bij eenvoorheen normotensieve zwangere. Dezelfde criteria worden gebruikt voor deidentificatie van preëclamptische zwangeren: pih wordt gecombineerd met pro-teinurie boven 300 mg/24 h of minimaal 2+ eiwit bij dipstick urine analyse na 20

weken zwangerschap. Intra-uteriene groeirestrictie wordt gedefinieerd als neo -nataal gewicht bij geboorte onder het tiende percentiel (gecorrigeerd voorzwangerschapsduur en geslacht).

In chapter 2 wordt de reproduceerbaarheid van driedimensionale echoscopische(3dus) metingen van foetale en placentaire volumes in zwangerschapsweek 11tot 18 in 34 ongecompliceerde zwangerschappen geëvalueerd.

Twee onderzoekers verzamelde onafhankelijk van elkaar foetale en placen-taire volumes met behulp van 3dus. Elk foetaal en placentair volume werd tweekeer gemeten. Intra- en interobserver reproduceerbaarheid werd uitgedrukt inintra- en interclass correlation coëfficiënt (intra-cc en inter-cc). Daarbij werdhet effect van de individuele volumemeting en caliperplaatsing onderzocht.

Foetale en placentaire volume metingen werden succesvol verricht in 97%van alle gevallen en beide volumes kwamen overeen met volumes uit eerderegepubliceerde studies. De intra-observer reproduceerbaarheid was goed voor

113CHAPTE R 9

zowel foetale volumemetingen (intra-cc: 0.99; 0.99) als placentaire volume -metingen (intra-cc: 0.99; 0.98). Daarbij bleef de inter-cc goed voor foetale (inter-cc: 0.99; 0.99) en placentaire volumemetingen (inter-cc: 0.97; 0.99) on-afhankelijk van de onderzoeker die de metingen verrichtte.

Onze conclusie is dat intra- en interobserver variabiliteit van foetale en pla-centaire volumemetingen laag is en de reproduceerbaarheid zeer goed. Daarbijis aangetoond dat er geen effect is als gevolg van individuele volumemeting ofcaliperplaatsing op de uiteindelijke foetale en placentaire volumemeting.

In chapter 3 wordt de klinische waarde van 3D us metingen van foetale en pla-centaire volumemetingen beschreven.

Driedimensionale echoscopische foetale volumemetingen zijn gebaseerd ophoofd- en rompvolumes, ze zijn zeer goed reproduceerbaar en lijken erg accu-raat ten aanzien van foetale gewichtsschatting en de evaluatie van foetale groei.Foetale volumemetingen zijn echter beperkt door zwangerschapsduur, daar degrootte van laat-tweede en derde trimester foetussen over het algemeen buitenhet bereik van de volumebox valt. De klinische waarde van eerste trimester volume -metingen moet nog bepaald worden, maar significant lagere volumes zijn aange-toond in aneuploïde foetussen in vergelijking met normale foetussen.

Ook placentaire volumemetingen zijn zeer goed reproduceerbaar en ge mak -kelijk te verrichten door middel van 3dus. Net als foetale volumemetingen, wordenook de placentaire volumemetingen beperkt door de grootte van de volumebox.Meerdere studies laten een significant verlaagd placentavolume zien bij aneu-ploïde foetussen, zoals trisomie 13, 18 en 21, in vergelijking met normale foe-tussen. Hoewel de voorspellende waarde voor foetale aneuploïdie beperkt is envooralsnog tot weinig verbetering in detectiegraad van de huidige screenings -onderzoeken oplevert. De relatie tussen tweede trimester placentaire volume -metingen en zwangerschappen gecompliceerd door hypertensieve stoornissenof foetale groeirestrictie lijkt veelbelovend, hoewel de voorspellende waardebeperkt lijkt door de grote heterogeniciteit in placenta groei en grootte. Meeronderzoek voor het bepalen van de voorspellende waardes van eerste trimesterfoetale en placentaire volumemetingen is nodig.

chapter 4 beschrijft de longitudinale relatie tussen Echo Doppler metingen vande uteriene arteriën (ua’s) en de spiraal arteriën (sa’s) in zwangere vrouwen. Indeze studie analyseerden we 97 primigravida’s met een ongecompliceerde een-lingzwangerschap. In elke zwangerschap werden de bloedstroomsnelheden vanbeide (gecombineerde) ua’s en twee sa’s gemeten door middel van transabdo -


mi naal echoscopisch Color Doppler onderzoek tussen zwangerschapsweek 11tot 14, 14 tot 18 en 18 tot 24. Elke Echo Dopplermeting werd tweemaal uitgevoerd.De gecombineerde ua werd uitgerekend door het gemiddelde te nemen van delinker- en de rechter uteriene arterie. Daarbij werd het voorkomen van bilate -rale notching genoteerd.

In totaal werden 284 ua en 263 sa Echo Doppler metingen geanalyseerd. Deresultaten lieten een continue afname van de gemiddelde pi in de ua’s en de sa’szien bij toenemende zwangerschapsduur. Een bilaterale notch was zichtbaar in35%, 9% en 3% van de cases tussen zwangerschapsweek 11 tot 14, 14 tot 18 en 18tot 24 respectievelijk. De intra-observer variabiliteit voor de gecombineerde ua’sen de sa’s was 0.90 en 0.54 respectievelijk. De Generalized Estimating Equationsanalyse liet een significante relatie (r = 0.41) zien tussen ua’s en sa’s (p < 0.0001).

Deze longitudinale data bevestigt de algemeen geaccepteerde relatie tussende uteriene en spiraal arteriën en laat een significante correlatie tussen beidemetingen zien: ua Echo Doppler metingen lijken een goede weergave te gevenvan de perifere weerstand door spiraal arteriën. Daarbij laat deze studie zien dat detrophoblastinvasie een continu proces is in de eerste helft van de zwangerschap,waarbij het voorkomen van bilaterale notches waarschijnlijk een fysiologischfenomeen is.

In chapter 5 evalueren we de longitudinale relatie tussen sa Echo Doppler me -tin gen en zwangerschappen gecompliceerd door pe, pih of iugr in 108 zwanger-schappen.

In alle zwangerschappen werden de sa bloedstroomsnelheden driemaal ge -meten door middel van Color Doppler. De metingen werden verricht tussenzwan ger schapsweek 11 tot 14, 14 tot 18, en 18 tot 24. Elke Echo Doppler metingwerd tweemaal uitgevoerd.

De sa Echo Doppler metingen werden geanalyseerd met behulp van multi-level analyse en referentiecurven werden berekend. De Mann-Whitney test werdgebruikt om de ongecompliceerde en gecompliceerde zwangerschappen te ver -gelijken.

Er werden 86 ongecompliceerde zwangerschappen en 21 gecompliceerdezwangerschappen (4 pe, 7 pih en 10 iugr) geanalyseerd. In de ongecompliceerdezwangerschappen daalden de systolic/diastolic (sd-)ratio’s, resistance (ri) en pul -satility indices (pi) progressief bij toenemende zwangerschapsduur. Een lineaireregressieanalyse liet een gemiddelde sd-ratio zien van 1.75 bij zwangerschaps -week 11+0 (P5 – P95: 1.32 – 2.17) tot 1.48 bij zwangerschapsweek 24+0 (P5 – P95:1.05 – 1.90). Een soortgelijke trend was te zien bij de ri en pi: de gemiddelde ri

115CHAPTE R 9 SAMENVATTING

daalde van 0.44 (P5 – P95: 0.28 – 0.59) bij zwangerschapsweek 11+0 naar 0.34

(P5-P95: 0.19 – 0.50) bij zwangerschapsweek 24+0, de gemiddelde pi van 0.59

(P5 – P95: 0.32 – 0.85) bij zwangerschapsweek 11+0 naar 0.40 (P5 – P95: 0.14 – 0.66)bij zwangerschapsweek 24+0 weeks. Deze resultaten kwamen overeen met eer -der gepubliceerde data. De Mann-Whitney test liet geen significante verschillenzien in sd-ratio, ri of pi in zwangerschappen gecompliceerd door pe of iugr. Erwaren een significant lagere sd-ratio, ri en pi in zwangerschappen gecompliceerddoor pih.

Deze resultaten laten zien dat sa Echo Doppler metingen progressief dalenmet toenemende zwangerschapsduur, waarbij de referentiewaarden breed zijn.Daarom lijken sa Echo Doppler metingen niet geschikt voor vroege voorspellingvan pe of iugr. Verlaagde Doppler metingen in zwangerschappen gecompliceerddoor pih lijken te worden veroorzaakt door een hyperdynamische circulatie invrouwen met een verhoogde bmi en (gemaskeerde) pre-existente hypertensie.

In chapter 6 evalueren we de maternale serum markers adam12s and pp13 alseerste trimester voorspellers van gecompliceerde zwangerschapsuitkomst, zoalspe, pih en iugr. Een retrospectieve case-control studie van maternale serumsam-ples werd hiervoor uitgevoerd, waarbij de zwangerschapsduur varieerde van9+0 tot 13+6 weken. Alle serumsamples werden verzameld als onderdeel vanhet downsyndroom screening programma van het vumc in de periode van 2004-2007. Alle pe, pih en iugr cases werden gematched voor zwangerschapsduur,primipariteit, etniciteit, roken, maternaal gewicht en leeftijd met 3 gezonde con-troles. De serumconcentraties van adam12s en pp13 werden bepaald zonder datde zwangerschapsuitkomst bekend was en uitgedrukt in multiples of the median(mom). mom-waardes en hun voorspellende waarde werden geanalyseerd doormiddel van Mann-Whitney testen en receiver-operator-characteristics (roc)curven.

Honderdvijfenzestig controlemonsters en 55 cases (17 pe cases, 30 pih casesen 8 iugr) cases werden geïncludeerd. Mediane adam12s concentraties voor decontroles versus de cases waren significant verlaagd: 405 versus 324 nG/L (mom

1.00 versus 0.80 (p < 0.05)). Bij pp13 waren geen significante verschillen tussende controles en de cases; 57.7 vs. 54.6 pG/L (mom 1.00 and 0.95).

Mediane mom levels voor adam12s waren 0.90, 0.77 en 0.88 voor pe, pih andiugr respectievelijk; mom levels voor pp13 waren 0.77, 0.95 en 0.89 respec-tievelijk. roc curven lieten een area under the curve (auc) zien voor adam12s enpp13 van 0.63 en 0.59 voor pe, 0.68 en 0.57 voor pih en 0.59 en 0.62 voor iugr,respectievelijk. Gecombineerde adam12 en pp13 bepalingen verhoogde de auc


niet. Bij een specificiteit van 0.80, was de bijbehorende sensitiviteit van adam12s52% voor pih.

Uit deze resultaten blijkt dat verlaagde eerste trimester levels van adam12sgeassocieerd zijn met pih. Verlaagde levels van pp13 waren niet significant gecor-releerd met pe, pih of iugr. Gecombineerde adam12s en pp13 bepalingen ver-beteren de voorspellende waardes niet.

In chapter 7 wordt de relatie tussen eerste trimester plasma concentraties vanMetastin en foetale groeirestrictie beschreven.

In deze studie werd maternaal plasma afgenomen tussen zwangerschaps -week 8 tot 14 en daarin werd de concentratie van metastin bepaald. Eenendertigzwangerschappen gecompliceerd door sga-foetussen werden gematched met 31

ongecompliceerde zwangerschappen voor zwangerschapsduur bij plasma afname.ß-hCG bepalingen werden ook verricht om het effect van zwangerschapsduur enplacentavolume op de concentratie metastin te beoordelen.

De resultaten lieten een significant lagere plasmaconcentratie metastin zien inzwangerschappen gecompliceerd door sga-foetussen vergeleken met ongecom -pliceerde zwangerschappen (metastin: 1376 ± 1317 pmol/L versus 2035 ± 1260

pmol/L, p = 0.035), ß-hCG concentraties waren niet significant verschillend.Concluderend zijn metastin bepalingen in het eerste trimester mogelijk bruik-

baar voor vroege voorspelling van foetale groeirestrictie, waarschijnlijk in combi-natie met andere maternale markers.

117CHAPTE R 9 SAMENVATTING

List of co-authors

prof dr MA BlankensteinDepartment of Clinical ChemistryVU University Medical CenterAmsterdam, the Netherlands

dr AC BolteDepartment of Obstetrics and GynecologyVU University Medical CenterAmsterdam, the Netherlands

dr ATJJ GoDepartment of Obstetrics and GynecologyVU University Medical CenterAmsterdam, the Netherlands

dr AC HeijboerDepartment of Clinical ChemistryVU University Medical CenterAmsterdam, the Netherlands

drs JW KistDepartment of Obstetrics and GynecologyLeiden University Medical CenterLeiden, the Netherlands

Drs DJ KuikDepartment of Clinical Epidemiology andBiostatistics, VU University Medical CenterAmsterdam, the Netherlands

drs IH LinskensDepartment of Obstetrics and GynecologyVU University Medical CenterAmsterdam, the Netherlands

dr CBM Oudejans

Department of Clinical ChemistryVU University Medical CenterAmsterdam, the Netherlands

Mw Rekoert-HollanderDepartment of Obstetrics and GynecologyVU University Medical CenterAmsterdam, the Netherlands

drs EML SmetsScience & Safety OfficerRoche Diagnostics bv

Almere, the Netherlands

drs MD SpreeuwenbergKenniscirkel Technologie in de zorgHogeschool ZuydHeerlen, the Netherlands

prof dr JWR TwiskDepartment of Clinical Epidemiology andBiostatisticsVU University Medical CenterAmsterdam, the Netherlands

prof dr JMG van VugtDepartment of Obstetrics and GynecologyVU University Medical CenterAmsterdam, the Netherlands

118

List of publications

1. Deurloo KL, Devlieger R, Oepkes D. Maternal hydrops syndrome following successfultreatment of fetal hydrops by shunting of bilateral hydrothorax. Prenatal Diagnosis 2003.

23 (11): 944-945

2 Deurloo KL, Cobben JM, Heins YM, de Ru M, Wijnaendts LCD, van Vugt JMG. Prenataldiagnosis of tetrasomy 9p in a 19 week old fetus with Dandy-Walker malformation: aCase Report. Prenatal Diagnosis 2004. 24 (10): 796-798

3. Deurloo KL, Devlieger R, Loprior E, Klumper FJ, Oepkes D. Isolated fetal hydrothoraxwith hydrops: a systematic review of prenatal treatment options. Prenatal Diagnosis 2007.

27 (10): 893-899

4. Deurloo KL, Spreeuwenberg MD, Rekoert-Hollander GCM, van Vugt JMG.Reproducibility of three-dimensional ultrasound measurements of fetal and placentalvolume in week 11–18 of gestation. J Clin Ultrasound 2007. 35 (3): 125-132

5. Deurloo KL, Spreeuwenberg MD, van Vugt JMG. Colour Doppler ultrasound of spiralartery blood flow for prediction of hypertensive disorders and intra uterine growthrestriction. Prenatal Diagnosis 2007. 27(11): 1011-1016

6. Deurloo KL, Kist WJ, van Vugt JMG. 3D Ultrasound: an overview of its clinical use. Fetal and Maternal Medicine Review 2007. 18 (2) 145-179

7. Smets EML, Deurloo KL, Go ATJJ, van Vugt JMG, Blankenstein MA, Oudejans CBM.Decreased plasma levels of metastin in early pregnancy are associated with small forgestational age neonates. Prenatal Diagnosis 2008. 28(4) 299-303.

8. Deurloo KL, Bolte AC, Spreeuwenberg MD, van Vugt JMG: Longitudinal Dopplermeasurements of spiral artery blood flow in relation to uterine artery blood flow. JUltrasound in Medicine 2009 28(12) 1623-1628

9. Deurloo KL, Linskens I, Heijboer AC, Blankenstein MA, van Vugt JMG. adam12s andpp13 as first trimester predictors of adverse pregnancy outcome. Revision AJOG

10. Keizer I, Deurloo KL, Boonstra A, Melgers I, Mijatovic V, Hompes P. Spontane pneu -mothorax bij vrouwen; denk aan thoracale endometriose. Modern Medicine 2010 (2) 4-7.

11. Zwijnenberg PJG, Deurloo KL, Waterham HG, Meijers-Heijboer EJ, van Vugt JMG, Tan-Sindunata MB. Second trimester prenatal diagnosis of rhizomelic chondrodysplasiapunctata type 1 on ultrasound findings. Prenatal Diagnosis 2010. 30 (2) 162-164

12. Dumoulin CM, van Leeuwen JA, Ket JCF, Deurloo KL, Mijatovic V. Is een Mirenaspiraaleen goede behandeloptie ten aanzien van dysmenorroe en hypermenorroe bijadenomyose? ntog 2010 123 (2) 71-73

119

Dankwoord

Promoveren lukt je niet alleen. Een aantal mensen wil ik in het bijzonder bedanken.

Deelneemsters aan mijn onderzoekVeel dank ben ik verschuldigd aan alle deelneemsters aan mijn onderzoek. Uwtijd, geduld en enthousiasme voor mijn onderzoek heb ik zeer op prijs gesteld,alsook uw wervend vermogen onder vrienden en familie voor deelname aanmijn onderzoek. Het was bijzonder om betrokken te zijn bij uw zwangerschap enuw verwachtingen en toekomstplannen te horen. De meerderheid onder u hadprobleemloze zwangerschappen met prachtig eindresultaat, waar ik u veel gelukmee wens. Enkelen van u maakten helaas een ernstige zwangerschapscomplicatiedoor; hun wil ik veel sterkte wensen en daarbij benadrukken hoe bijzonder ik het vond dat u toch betrokken bleef bij mijn onderzoek.

prof dr JMG van VugtBeste John, allereerst ben ik je zeer dankbaar voor de kans om te werken alsarts-echoscopist en arts-onderzoeker op jouw afdeling prenatale diagnostiek.Tijdens de sollicitatie maakte ik al kennis met je humor: ik vertelde over een review dat ik aan het schrijven was, waarop jij geïnteresseerd informeerde of ikjouw artikelen daarover ook gelezen had. Gelukkig was ik eerlijk, je had immersgeen enkel artikel over dat onderwerp gepubliceerd.

Het was het begin van een zeer leerzame periode op de afdeling prenatale diagnostiek, waarbij ik me volledig kon uitleven op mijn passie: de obstetrischeechoscopie. Tijdens de vele besprekingen en overlegmomenten heeft je kennisover de gezonde en zieke foetus altijd veel indruk op me gemaakt, alsook de mogelijkheid tot het laagdrempelig raadplegen van deze kennis. Ook de humorbleef, je maakte van mij een taalpurist die woorden als ‘groeivertraging’ of‘groeiretardatie’ verafschuwde (en terecht).De samenwerking op wetenschappelijk gebied verliep eveneens uitstekend. Jebent een geweldige promotor: stimulerend, vol ideeën en je leek wel continu online, zo snel kreeg ik mijn gecorrigeerde manuscripten en presentaties terug.Je liet me mijn promotie in mijn eigen tempo uitwerken, en ook als het tempodoor gezinsvorming en start van de opleiding wat lager lag, bleef je vertrouwenin de goede afloop. Iets waardoor ik me altijd gesteund heb gevoeld.

120

De usa heeft ons beider passie en op meerdere congressen aldaar hebben weons altijd uitstekend vermaakt. Met name Washington met zijn nachtelijke ‘foggybottoms’ was een absolute topper.

Bedankt voor alles, John. Ik ben vereerd jouw promovendus te zijn.

prof dr MA BlankensteinBeste Rien. Gids in het land van de klinisch-chemische dingen, waar ik me nogaleens onwerkelijk voelde. Gelukkig werden de vrijdagbesprekingen door jou ondertiteld met lekentaal. Het was buitengewoon interessant om ook andereklinische onderzoeken over voorspellende serum- en plasmamarkers te volgen.Dank voor de introductie in deze wereld van serummarkers en hun veelbelovendepotentie. Bedankt voor de fijne samenwerking.

dr AC BolteBeste Annemieke, naast het geven van een goede ‘opvoeding’ in de derdelijns-verloskunde, heb je mij met jouw encyclopedische kennis en ongekende vertrouwen in mijn data meerdere malen van de wanhoop gered. Door jouwcreativiteit kreeg mijn onderzoek een nieuwe dimensie, waarvoor ik je erg dankbaar ben. Daarnaast heb ik jou als een betrokken en gezellig persoon lerenkennen, waarbij ik vooral denk aan onze congresbezoeken in Miami en Chicago.Met name in Miami bleek jouw uitgebreide internationale netwerk een aantalgoede feestgangers te bevatten.Het is een eer om jouw eerste promovendus te zijn.

prof dr JWR Twisk, drs MD Spreeuwenberg, drs DJ KuikBeste Jos, Marieke en Joop. Het zal voor jullie niet altijd makkelijk zijn geweestom al mijn vragen en gedachten om te zetten in statistische analyses en conclusies, zeker gezien mijn statistische basiskennis. Gelukkig is het jullie altijdgelukt.Dank Marieke, voor het urenlang skypen om de analyses af te krijgen. Dank Joop,dat je naast de analyses ook nog tijd had om mij privécolleges over statistiek tegeven. Dank Jos, dat je altijd tussendoor tijd vond om me te helpen met de multi-level-analyses.

dr D OepkesBeste Dick, mijn eerste stappen in de wetenschap, de obstetrische echoscopieen de prenatale diagnostiek heb ik onder jouw begeleiding gezet. Dankzij jou

121

maakte ik een vliegende start in Amsterdam. Ik zie het als een mooie afsluitingvan mijn promotie dat je zitting hebt genomen in de leescommissie.

Collega-arts-echoscopistenBeste Melanie (de primus inter pares), Franca, Mireille, Lucas, Yolanda en Ingeborg, bedankt voor de samenwerking en hulp bij de inclusies. Ingeborg, bedankt voor je enorme hulp bij het uitzoeken van geschikte sera en de bijbehorende analyses. Succes met de afronding van jouw promotie. Lucas, er zijn niet veel gynaecologen die een 3dus-foto van hun eigen gezicht hebben,bedankt voor heel veel humor en succes met jouw promotie.

EchoscopistenCharlotte, Bridget, Fayette, Annemieke, Annelieke en Marijke. Veel van mijnechoscopische basisvaardigheden heb ik van jullie geleerd. Bedankt voor de immer goede samenwerking en uiteraard jullie flexibiliteit op de echokamers,waardoor mijn inclusieperiode niet te lang ging duren. Speciale dank gaat uit naar Marijke vanwege je enorme inzet bij de reproduceerbaarheidstudie (en de bijbehorende borrels met je man).

Klinisch chemici en laboratoriummedewerkersCees Oudejans en collega's, bedankt voor de serumanalyses en jullie geduldigeuitleg over de vaak ingewikkelde methodes. Speciale dank aan Carla Beersen voorje grote hulp bij de serumselectie en bijbehorende bepalingen.

drs E SmetsBeste Eva, zeer dankbaar ben ik voor je aanzienlijke bijdrage aan mijn proefschrift,met name met betrekking tot de serummarkers. Ik heb genoten van je humor, jecharme en onze samenwerking op het laboratorium.

Collega-arts-onderzoekersBianca, Marja, Majoie, Marinka, Annelies, Marieke V, Tatjana, Iris, Marja Liisa, Esther, Sjanneke, Marieke L, Lucas, Yolanda, Ingeborg en Mireille. Bedankt voorde gezellige tijd op de Meander en de etentjes bij immer goed bereikbare restaurants.

dr MC HaakBeste Monique, dank voor de vele tussendooroverlegjes over abstracts, artikelen,presentaties, proefschriften en meer. Het kon altijd en je was razendsnel met

122

correcties. Met name je creativiteit om binnen de maximum wordcount van eenabstract te blijven, is indrukwekkend.

dr AGJJ GoBeste Atti, ik zal je combinatie van cola light, Leiden en humor nog missen. Bedanktvoor je hulp bij het opzetten van de Acces-database. Veel succes in Rotterdam.

Secretariaat Prenatale DiagnostiekBeste Lucia, Marijke, Marleen en Ineke. Dank voor alle gezelligheid, hulp bij de follow-up, de interessante gesprekken en de vele ‘echte’ koffie.

Gynaecologen en alle (oud-)collega-assistenten uit het VU Medisch CentrumBeste allen, het zit er eindelijk op. Dank voor jullie interesse in het verloop en hetuiteindelijke resultaat van mijn studie.

Paranimfen,

dr MN BekkerLieve Mireille. Samen startend aan onze promotietrajecten als arts-echoscopistin het vumc vonden we elkaar snel. Je was een erg relaxte kamergenoot en zeerloyale collega. Goede herinneringen bewaar ik aan de vele ‘IVF-besprekingen’met Diederik, Hans en Franyke. Gedurende de jaren werden we van collega’sgoede vrienden en hebben we ons op veel cursussen en congressen altijd bijzonder vermaakt. De relaxte borrel aan Ocean Drive vat alles goed samen,denk ik. Gelukkig weten we in deze drukke tijd van opleiding en gezinsuitbreidingelkaar nog regelmatig te vinden voor een avond chillen. Het is niet meer dan logisch dat jij mijn paranimf bent.

Renzo DeurlooMijn lieve broer, ofwel Bro. Met jou als paranimf neem ik het risico dat je de verdediging van mijn proefschrift ongevraagd overneemt... en waarschijnlijk metde jou kenmerkende charme ook nog succesvol weet af te ronden. Met je positieve instelling en gevoel voor humor ben je een steun in veel dingen, zo ookdit proefschrift. Ik ben er trots op dat je naast me staat tijdens de verdediging.

GriffioenOmdat het kan.

123

Vrienden, familie en schoonoudersBedankt voor de hulp op vele manieren, die het mogelijk maakte mijn promotieaf te ronden. Speciale dank voor mijn schoonouders om hun voordurende interesse en steun en voor mijn lieve zusje Sandra om haar includerend vermogenen kunst van het relativeren.

Renée DeurlooLief zusje. Gelukkig herinnerde je me tijdig aan je beroep als redactrice, anderszat ik zelf nog te knutselen in Word Perfect. Ik heb genoten van onze samen -werking en ben je erg dankbaar voor alle uren die je hebt besteed aan de correctieen verbetering van mijn proefschrift.

Pa en MaBedankt voor de fantastische vooropleiding en jullie onvoorwaardelijke liefde en vertrouwen. Jullie zijn de besten.

SaskiaAllerliefste Guapa, mijn highschool sweetheart, onmisbaar in zo veel dingen...Heel veel dank voor heel veel moois.

Madelief en OlijfMijn meisjes, mooier en liever dan jullie worden ze niet meer gemaakt.

early prediction of preeclampsia and iugr dissertation.pdf · early prediction of preeclampsia and...

Documents