Uterine Artery Doppler ISUOG 2013

7/29/2019 Uterine Artery Doppler ISUOG 2013

1/11

Ultrasound Obstet Gynecol2013; 42: 257267Published online 6 August 2013 in Wiley Online Library (wileyonlinelibrary.com). DOI: 10.1002/uog.12435

Value of adding second-trimester uterine artery Doppler topatient characteristics in identification of nulliparous womenat increased risk for pre-eclampsia: an individual patientdata meta-analysis

C. E. KLEINROUWELER*, P. M. M. BOSSUYT, B. THILAGANATHAN, K. C. VOLLEBREGT*,J. ARENAS RAMIREZ, A. OHKUCHI, K. L. DEURLOO**, M. MACLEOD, A. E. DIAB,H. WOLF*, J. A. M. VAN DER POST*, B. W. J. MOL* and E. PAJKRT*

*Department of Obstetrics and Gynaecology, Academic Medical Center, Amsterdam, The Netherlands; Department of ClinicalEpidemiology, Biostatistics and Bioinformatics, Academic Medical Center, Amsterdam, The Netherlands; St Georges Healthcare NHSTrust, Department of Obstetrics and Gynaecology, London, UK; Hospital de Cabuenes, Department of Obstetrics and Gynaecology,Gijon, Spain; Jichi Medical University School of Medicine, Department of Obstetrics and Gynecology, Shimotsuke-shi, Tochigi-ken, Japan;**VU University Medical Center, Department of Obstetrics and Gynaecology, Amsterdam, The Netherlands; University of Dundee,Division of Clinical & Population Sciences & Education, Academic Clinical Practice, Ninewells Hospital & Medical School, Dundee, UK;Zagazig Faculty of Medicine, Department of Obstetrics and Gynaecology, Zagazig, Ash Sharqiyah, Egypt

K E Y W O R D S : Doppler ultrasound; individual patient data meta-analysis; prediction; pre-eclampsia; uterine artery

ABSTRACT

Objective To investigate the value of adding second-trimester uterine artery Doppler ultrasound to patientcharacteristics in the identification of nulliparous womenat risk for pre-eclampsia.

Methods For this individual patient data meta-analysis,studies published between January 1995 and December2009 were identified in MEDLINE and EMBASE. Studieswere eligible in which Doppler assessment of the uterinearteries had been performed among pregnant womenand in which gestational age at ultrasound, Dopplerultrasound findings and data on the occurrence of

pre-eclampsia were available. We invited correspondingauthors to share their original datasets. Data wereincluded of nulliparous women who had had a second-trimester uterine artery Doppler ultrasound examination.Shared data were checked for consistency, recoded toacquire uniformity and merged into a single dataset. We

constructed random intercept logistic regression modelsfor each of the patient and Doppler characteristics inisolation and for combinations. We compared goodnessof fit, discrimination and calibration.

Results We analyzed eight datasets, reporting on 6708nulliparous women, of whom 302 (4.5%) developed pre-eclampsia. Doppler findings included higher, lower andmean pulsatility index (PI) and resistance index (RI) and

Correspondence to: Ms E. Kleinrouweler, Academic Medical Center, Department of Obstetrics and Gynaecology, Room H4-232,

Meibergdreef 9, 1105 AZ Amsterdam, The Netherlands (e-mail: [email protected])

Accepted: 1 February 2013

any or bilateral notching. Of these, the best predictorswere combinations of mean PI or RI and bilateralnotching, with areas under the receiveroperatingcharacteristics curve (AUC) of 0.75 (95% confidenceinterval (CI), 0.560.95) and 0.70 (95% CI, 0.66 0.74),respectively. Addition of Doppler findings to the patientcharacteristics blood pressure or body mass index (BMI)significantly improved discrimination. A model withblood pressure, PI and bilateral notching had an AUCof 0.85 (95% CI, 0.671.00).

Conclusions The addition of Doppler characteristicsof mean PI or RI and bilateral notching to patientcharacteristics of blood pressure or BMI improves theidentification of nulliparous women at risk for pre-eclampsia. Copyright 2013 ISUOG. Published by

John Wiley & Sons Ltd.

INTRODUCTION

Pre-eclampsia is a multisystem disorder of pregnancy,defined as the presence of hypertension and proteinuria,that contributes substantially to maternal and perinatalmorbidity and mortality worldwide1. The obstetricmanagement of women with pre-eclampsia is complicatedby the fact that by the time symptoms occur, theonly definitive treatment of the underlying disorder

Copyright 2013 ISUOG. Published by John Wiley & Sons Ltd. META-ANALYSIS


2/11

258 Kleinrouweler et al.

is delivery. Management focuses on safe prolongationof pregnancy through intensive monitoring, such thatmaternal and fetal complications can be prevented1.Accurate prediction of this condition in early pregnancywould allow for timelyallocation of monitoring resources,with the prospect of improving maternal and perinataloutcomes2,3. Preventive treatment such as aspirin seems

more likely to be beneficial if started earlier inpregnancy4.

Vascular supply to the placenta depends on trophoblastinvasion and remodeling of the spiral arteries duringthe late first and early second trimesters. A deficiencyin this process is associated with fetal growth restrictionand the development of pre-eclampsia. Doppler ultra-sound can be used to assess blood flow velocity in thematernal uterine arteries and thus potentially identifypregnancies at increased risk for pre-eclampsia and,indeed, numerous studies have investigated screeningstrategies for pre-eclampsia based on uterine arteryDoppler ultrasound. Combination screening strategiesbased on patient characteristics and multiple markers ortests, such as Doppler ultrasound, seem more promisingthan tests in isolation511.

There is uncertainty about the prognostic accuracyof Doppler ultrasound findings, when combined withmore readily available patient characteristics suchas blood pressure and being overweight. In clinicalpractice, a test will always be performed when thereis prior suspicion, the strength of which will varybetween patients. The question is not whether Dopplerultrasound in isolation can distinguish between futurepre-eclampsia or uncomplicated pregnancy, but whether

Doppler ultrasound results add relevant information towhat clinicians have already inferred from the clinicalcharacteristics of the pregnant woman.

Conventional meta-analyses of prognostic studies onlysummarize data on the accuracy of a single test underconsideration and do not make use of all availableinformation. In contrast, individual patient data (IPD)meta-analysis allows one to compare multivariableprediction strategies and also allows the possibility oftime-to-event analysis12.

We report here on a project to investigate the addedvalue of uterine artery Doppler measurements in the

identification of nulliparous women at risk for pre-eclampsia, based on analyses of individual patient datafrom previously published studies.

METHODS

The IPD-POPULAR project (Prediction Of Pre-eclampsiafrom Doppler ULtrasonography, Anthropometric param-eters and maternal Risk factors) relies on a systematicsearch of the literature, invitations to share data and acomparison of multivariable prediction models for pre-eclampsia in these data. Under Dutch law, this projectdid not require formal approval of an ethics committeeor institutional review board, as was confirmed by theMedical Ethics Committee of the Academic Medical

Center. This meta-analysis was conducted according tothe Meta-analysis Of Observational Studies in Epidemi-ology (MOOSE) guidelines13.

Literature search, study selection, data collectionand quality assessment

We searched MEDLINE and EMBASE between 1995and 2009 to identify eligible studies. Studies were eligibleif Doppler assessment of the uterine arteries had beenperformed in pregnant women, at any gestational age,in any healthcare setting, at any level of risk for pre-eclampsia and in which gestational age at ultrasound,Doppler ultrasound findings and the occurrence ofpre-eclampsia had been recorded. Acceptable referencestandards for pre-eclampsia were persistent high systolic(140mmHg) or diastolic (90 mmHg) blood pressureand proteinuria (0.3 g/24 h or a dipstick result of1+,equivalent to 30mg/dL in a single urine sample) of

new onset after 20 weeks gestation, according to theInternational Society for the Study of Hypertensionin Pregnancy criteria14. More details of the searchstrategy and selection of eligible studies can be found inAppendix S1 online.

Over a period of 3 years (20082010), the corres-ponding authors of eligible studies were contacted by e-mail and invited to participate in the project and sharetheir original datasets, i.e. those that had been used forthe analyses published in the study reports. We onlycontacted authors of studies published from 1995, as weassumed that data collected earlier would possibly notbe stored in an electronic format still useful today, and

because Doppler ultrasoundtechniqueshavechanged overtime. When authors replied positively they were providedwith a detailed project protocol and were asked to sendtheir original, complete and anonymized data. Additionalinformation was extracted from the original papers (seealso Appendix S1).

The methodological quality of the studies wasassessed from the original publication using the QualityAssessment of Diagnostic Accuracy Studies (QUADAS)criteria15. In addition, we checked for consistencybetween received data and the results in the publishedpaper. Shared data were reformatted or recoded if

necessary to ensure uniformity, and merged into a singledataset.

Data analysis

For this analysis16,17, we restricted ourselves to nulli-parous women who had had a second-trimester uterineartery Doppler ultrasound examination. The relation-ship between each patient characteristic or Dopplerultrasound parameter and pre-eclampsia was evalu-ated by univariable logistic regression analysis. Wemodeled continuous variables with quadratic, cubicand logarithmic terms to assess potential non-linearity.Two-level-random intercept logistic regression modelswere then constructed for each of the available Doppler

Copyright 2013 ISUOG. Published by John Wiley & Sons Ltd. Ultrasound Obstet Gynecol2013; 42: 257267.


3/11

IPD meta-analysis of UtA for PE 259

parameters in isolation and in combinations, using the

occurrence of pre-eclampsia at the individual patient

level as the outcome variable. Stratification at the study

level was used to control for remaining between-study

heterogeneity18.

We identified the best predictive patient characteristic

or combination of patient characteristics, andalso thebest

predictive Doppler parameter, or combination of Dopplerparameters, in terms of model fit. Then we compared

discrimination with a model consisting of only patient

characteristics, a model with Doppler parameters only,

and a model containing both patient characteristics and

Doppler ultrasound findings.

Improvement in model fit was evaluated using the like-

lihood ratio test statistic for nested submodels (comparing

models with one or more additional variables). For non-

nested submodels, testing whether one submodel was

better than the other was done by calculating the differ-

ence in log likelihood between the two models in 10000

bootstrap samples drawn from the original dataset. Ineach bootstrap, as many observations were drawn from

each shared dataset as there were women included in that

dataset. P-values were calculated assuming a normal dis-

tribution of the differences in log likelihood with a true

mean of 0 and the observed standard deviation in the

10 000 bootstrap samples.

For the univariable analyses of the patient charac-

teristics and Doppler parameters we used data of all

the women in whom that variable had been recorded.

For the analyses in which variables were compared or

evaluated in combination, we used data of a subset of

all women for whom all studied variables were avail-able. As not all factors were present in all datasets, the

numbers of women studied in each step of the analysis

differed.

Model discrimination was assessed by calculating the

area under the curve (AUC) in a receiveroperating

characteristics (ROC) plot, also known as the c-statistic.

Differences in AUCs between models were tested with the

comproc function in R, which estimates the difference

between model AUCs in 10 000 bootstrap samples and

uses the Wald test for model comparisons based on the

bootstrap standard errors. Calibration of the models was

presented in calibration plots.The analyses were repeated with pre-eclampsia requir-

ing delivery at p75p80,

>p80p85, >p85p90, >p90p95 and>p95) and

women with probabilities of p75, according to the

model containingbothpatient characteristicsand Doppler

parameters, and expressed the results in KaplanMeier

plots. All analyses were performed using R version 2.12.1

(R Foundation, Vienna, Austria) and PASW statistics 18

(IBM Inc., New York, NY, USA).

RESULTS

Study selection and data collection

The searches in MEDLINE and EMBASE (1995 to 2009)resulted in a total of 3199 citations after removal ofduplicates, of which 176 study reports, written by 111different corresponding authors, were deemed eligible for

inclusion. We were able to contact 107 of them, of whom49 (46%) replied that they were interested in the projectand willing to share data. Twenty-two authors did notshare data despite an expressed intention to do so (n=17)or informed us that the data were no longer available(n=3) or that they had not been given institutional reviewboard approval for data sharing (n=2). Eventually, 27authors shared their datasets of 30 studies.

From the shared datasets, we were able to use eight forthis project, as these had data on nulliparous womenwho had had a second-trimester Doppler ultrasoundexamination1926. In total, data were available on 6708

unselected nulliparous women, of whom 302 (4.5%)developed pre-eclampsia. Table 1 summarizes studydetails, baseline patient characteristics and additionalDoppler measurements in these datasets. Table S1 (inAppendix S1 online) describes the study groups in thedatasets included in our analyses, and shows the numberof nulliparous women in the original studies. The resultsof the quality assessment are described in Appendix S1.Evaluation with the QUADAS instrument showed overallgood study quality.

Table 2 illustrates the availability of the variables ineach included dataset. As expected, not all variables wereavailable in all datasets. The results of the investigation

and selection of patient characteristics can also be foundin Appendix S1. We found that body mass index (BMI)and systolic blood pressure contributed most to theprediction of pre-eclampsia among the available patientcharacteristics. The combination of BMI and systolicblood pressure predicted pre-eclampsia better than BMIalone, but as well as systolic blood pressure alone.

Selection of Doppler ultrasound predictors

In five datasets combined, containing data from 3116women, we were able to investigate the relationship

between the higher, lower and mean (of the measuredleft and right uterine artery) pulsatility index (PI). Eachwas significantly associated with pre-eclampsia. Of thesethree, the mean PI had the best predictive performance,but this was not significantly better than that of the lowerPI (P=0.57) or higher PI (P=0.06).

The same was found to hold for higher, lower and meanresistance index (RI). In four combined datasets with datafrom 6271 women, all were significantly associated withpre-eclampsia, but the mean RI had the best predictiveperformance. Mean RI performedsignificantly better thanlower RI (P=0.005) but not significantly better thanhigher RI (P=0.33).

The prognostic value of bilateral and any notchingcould be assessed in six combined datasets containing



4/11


Table

1

Characteristicsoftheeightdatasetsand

includednulliparouswomenwhounderw

entsecond-trimesteruterineartery(UtA)

Dopplerultrasoundexamination

Study

Param

eter

1

2

3

4

5

6

7

8

Correspondingauthor

Arenas

Deurloo

Diab

Macleod

Ohkuchi

Thilaganathan

Vollebregt

Wolf

Country

Spain

TheNetherlands

Egypt

UK

Japan

UK

TheNetherlan

ds

TheNetherlands

Inclusionyears

20002001

Notreported

20052006

20012003

19931998

19962006

20042006

Notreported

Numb

erofwomen

193

123

32

100

146

5835

258

21

Maternalage(years)

29(1643)

32(2041)

22(1827)

NA

27(2042)

31(1555)

30(1842)

29(2437)

BMI(kg/m2)

24.0(17.325.8)

22.1(17.730.1)

22.0(19.028.0)

24.6(18.046.1)

NA

22.7(11.857.6)

22.7(17.941

.2)

21.6(18.730.9)

Smoke

r

60(31.1)

NA

0(0.0)

24(24.0)

NA

457(11.8)

NA

0(0.0)

SBP(m

mHg)

110(90159)

NA

NA

110(90140)

NA

NA

NA

112(102140)

GAat

ultrasound

(weeks)

20+

1(18+

1to

22+

3)

20+

1(17+

6to

24+

4)

23

1820

20+

2(16+

0to

23+

6)

21+

3(18+

0to

24+

0)

19+

3(16+6

to

28+

3)

20

UtAD

oppler

Mea

nPI

0.80(0.112.21)

0.91(0.372.05)

1.70(1.452.34)*

NA*

NA

0.80(0.272.89)

0.90(0.324.40)

0.83(0.572.22)

Mea

nRI

0.52(0.330.83)

0.56(0.310.84)

NA*

NA*

0.55(0.380.81)

0.53(0.231.32)

NA

NA

Bilateralnotching

17(8.8)

4(3.3)

23(71.9)*

44(44.0)*

21(14.4)

524(9.0)

34(13.4)

3(14.3)

AllPE

10(5.2)

5(4.1)

15(46.9)

14(14.0)

6(4.1)

238(4.1)

13(5.1)

1(4.8)

PEreq

uiring

delivery34weeks

1(0.5)

NA

5(15.6)

4(4.0)

0(0.0)

33(0.6)

4(1.6)

0(0.0)

PEreq

uiring

delivery37weeks

7(3.6)

NA

14(43.8)

7(7.0)

2(1.4)

71(1.2)

8(3.1)

0(0.0)

Usedforpatientfactor

mod

el

Yes

Yes

Yes

Yes

Yes

Yes

Yes

Yes

UsedforDoppler

mod

el

Yes

Yes

No*

No*

Yes

Yes

Yes

Yes

Usedforcombined

mod

el

Yes

Yes

No*

No*

Yes

Yes

Yes

Yes

Dataa

representedasn(%)ormedian(range).*WomeninthesestudieswereexcludedfromanalysesofDopplerultrasoundmeasurements,astheywereselectedbasedonab

normalDoppler

findings.Dataonsmokingavailablein3874ca

ses.BMI,bodymassindex;GA,gestationalage;NA,notavailableinthisdataset;P

E,pre-eclampsia;PI,pulsatilityindex;RI,

resistanceindex;

SBP,systolicbloodpressure.



5/11


Table 2 Availability of patient characteristics and Doppler ultrasound measurements per dataset included in this study

Study

Parameter 1 2 3 4 5 6 7 8 Total

Corresponding author Arenas Deurloo Diab Macleod Ohkuchi Thilaganathan Vollebregt Wolf

Nulliparous women 193 123 32 100 146 5835 258 21 6708

(100)

Age 193 122 32 0 146 5833 258 21 6605(100) (99.2) (100) (0) (100) (99.97) (100) (100) (98.5)

Height 193 0 0 100 0 4381 258 21 4953

(100) (0) (0) (100) (0) (75.1) (100) (100) (73.8)Weight 193 0 0 100 0 4326 258 21 4898

(100) (0) (0) (100) (0) (74.1) (100) (100) (73)

BMI 193 115 32 100 0 4249 258 21 4968(100) (93.5) (100) (100) (0) (72.8) (100) (100) (74.1)

Obesity (BMI30 kg/m2) 193 115 32 100 0 4249 258 21 4968

(100) (93.5) (100) (100) (0) (72.8) (100) (100) (74.1)

Smoker 193 0 32 100 0 3874 0 21 4220(100) (0) (100) (100) (0) (66.4) (0) (100) (62.9)

Alcohol consumption 0 0 0 0 0 3707 0 0 3707

(0) (0) (0) (0) (0) (63.5) (0) (0) (55.3)Ethnicity 0 0 0 0 146 5305 258 0 5709

(0) (0) (0) (0) (100) (90.9) (100) (0) (85.1)

SBP 193 0 0 99 0 0 0 21 313(100) (0) (0) (99) (0) (0) (0) (100) (4.7)

DBP 193 0 32 99 0 0 0 21 345

(100) (0) (100) (99) (0) (0) (0) (100) (5.1)

MAP 193 0 0 99 0 0 0 21 313(100) (0) (0) (99) (0) (0) (0) (100) (4.7)

Lower PI 193 97 0 0 0 2572 234 21 3117

(100) (78.9) (0)* (0)* (0) (44.1) (90.7) (100) (46.5)Higher PI 193 97 0 0 0 2572 234 21 3117

(100) (78.9) (0)* (0)* (0) (44.1) (90.7) (100) (46.5)

Mean PI 193 97 32 0 0 2572 234 21 3149(100) (78.9) (100)* (0)* (0) (44.1) (90.7) (100) (46.9)

Lower RI 193 97 0 0 146 5835 0 0 6271

(100) (78.9) (0)* (0)* (100) (100) (0) (0) (93.5)Higher RI 193 97 0 0 146 5835 0 0 6271

(100) (78.9) (0)* (0)* (100) (100) (0) (0) (93.5)

Mean RI 193 97 0 0 146 5835 0 0 6271

(100) (78.9) (0)* (0)* (100) (100) (0) (0) (93.5)Bilateral notching 193 122 32 100 146 5835 253 21 6702

(100) (99.2) (100)* (100)* (100) (100) (98.1) (100) (99.9)

Any notching 193 122 0 0 146 5835 241 21 6558

(100) (99.2) (0)* (0)* (100) (100) (93.4) (100) (97.8)Pre-eclampsia 193 123 32 100 146 5835 257 21 6707

(100) (100) (100) (100) (100) (100) (99.6) (100) (100)

Data are presented as n (%). *Women in these studies were excluded from analyses of Doppler ultrasound measurements, as they wereselected based on abnormal Doppler findings. BMI, body mass index; DBP, diastolic blood pressure; MAP, mean arterial pressure;PI, pulsatility index; RI, resistance index; SBP, systolic blood pressure.

data from 6569 and 6557 women, respectively. Bothbilateral and any notching were significantly associatedwith pre-eclampsia (both P


6/11


Table 3 Calculation of probabilities of pre-eclampsia from various predictors and combinations of predictors, including second-trimesteruterine artery Doppler

Predictors n Probability of pre-eclampsia

SBP 313 1/(1 + EXP((7.28164 + (0.04192 SBP))))BMI 4967 1/(1 + EXP((9.123706 + (0.404313 BMI) (0.005169 BMI2))))

SBP + BMI 313 1/(1 + EXP((3.766467 (0.295400 BMI) + (0.005813 BMI2)+ (0.043142 SBP))))

PI + bilateral notching 3116 1/(1 + EXP ((2.7260 + (0.5991 bilatnotch) (4.2688 PI)+ (4.5923 PI2) (1.0550 PI3))))

RI + bilateral notching 6271 1/(1 + EXP((12.4237 + (0.4714 bilatnotch) (89.5798 RI)+ (154.5781 RI2) (80.5996 RI3))))

SBP + PI + bilateral notching 214 1/(1 + EXP((1.72198 + (0.04097 SBP) (29.85228 PI)+ (25.86898 PI2) (6.50349 PI3) + (1.93245 bilatnotch))))

SBP + RI + bilateral notching 193 1/(1 + EXP((30.75249 + (0.02442 SBP) (192.07371 RI)+ (312.40119 RI2) (158.64051 RI3) + (1.58368 bilatnotch))))

BMI + PI + bilateral notching 2963 1/(1 + EXP((9.454603 + (0.446135 BMI) (0.006185 BMI2)+ (0.487445 bilatnotch) (5.209416 PI) + (5.375934 PI2) (1.241653 PI3))))

BMI + RI + bilateral notching 4533 1/(1 + EXP(( 8.576 + (0.4140 BMI) (0.005355 BMI2)+ (0.3521 bilatnotch) (104.6 RI) + (178.5 RI2) (92.20 RI3))))

SBP + BMI + PI + bilateral notching 214 1/(1 + EXP((0.8684 + (0.05955 BMI) (0.0004479 BMI2)

+ (0.03792 SBP) + (1.967 bilatnotch) (29.97 PI) + (26.06 PI2

) (6.567 PI3))))

SBP + BMI + RI + bilateral notching 193 1/(1 + EXP((21.86637 + (0.68734 BMI) (0.01232 BMI2)+ (0.02220 SBP) + (1.66428 bilatnotch) (191.65240 RI)+ (308.49609 RI2) (154.51456 RI3))))

n indicates number of women available for estimation of regression models, which was not always equal to number of women available forcomparison of multiple models as shown in Figures 13. EXP(x) = ex. bilatnotch, bilateral notching (yes/no); BMI, body mass index (kg/m2);PI, mean pulsatility index; RI, mean resistance index; SBP, systolic blood pressure (mmHg).

At this point in the analyses, we concluded that thebest models with Doppler characteristics included meanPI and bilateral notching, or mean RI and bilateralnotching. In the next step, we investigated the added

value of both combinations to the previously selectedpatient characteristics.

Added value of Doppler ultrasound measurementsto patient characteristics

As data on all patient characteristics and Dopplerparameters were not available for all the women, weinvestigated the added value of mean PI and bilateralnotching or mean RI and bilateral notching to BMI orsystolic blood pressure in isolation or in combination.The probabilities of pre-eclampsia were calculated for

individual patients as illustrated in Table 3.Figures 13 show the discriminative ability for all

combinations made. For all three combinations, thediscriminative ability of the model including both patientcharacteristics and Doppler parameters was better thanthat of a model with patient characteristics or Doppleralone. The AUC of the model containing systolic bloodpressure, mean PI and bilateral notching was 0.85 (95%confidence interval (CI), 0.67 1.00), higher than forthe model with systolic blood pressure only (AUC 0.64(95% CI, 0.450.84), P


7/11


1.000.00

0.10

0.20

0.30

0.40

0.50

Sensitivity 0.60

0.70

0.80

0.90

1.00(a)

0.90 0.80 0.70 0.60 0.50

Specificity

0.40 0.30 0.20 0.10 0.00 1.000.00

0.10

0.20

0.30

0.40

0.50

Sens

itivity 0.60

0.70

0.80

0.90

1.00(b)

0.90 0.80 0.70 0.60 0.50

Specificity

0.40 0.30 0.20 0.10 0.00

Figure 1 Receiveroperating characteristics curves showing discriminative abilities of systolic blood pressure (SBP), uterine artery Dopplerultrasound measurements and their combinations in the identification of women at risk for pre-eclampsia. (a) SBP combined with meanpulsatility index (PI) and bilateral notching. , SBP; , mean PI and bilateral notching; , SBP, mean PI and bilateral notching.

(b) SBP combined with mean resistance index (RI) and bilateral notching. , SBP; , mean RI and bilateral notching; , SBP,mean RI and bilateral notching.

1.000.00

0.10

0.20

0.30

0.40

0.50

Sensitivity 0.60

0.70

0.80

0.90

1.00(a)

0.90 0.80 0.70 0.60 0.50

Specificity

0.40 0.30 0.20 0.10 0.00 1.000.00

0.10

0.200.30

0.40

0.50

Sensitivity

0.60

0.70

0.80

0.90

1.00(b)

0.90 0.80 0.70 0.60 0.50

Specificity

0.40 0.30 0.20 0.10 0.00

Figure 2 Receiveroperating characteristics curves showing discriminative abilities of body mass index (BMI), uterine artery Dopplerultrasound measurements and their combinations in the identification of women at risk for pre-eclampsia. (a) BMI combined with meanpulsatility index (PI) and bilateral notching. , BMI; , mean PI and bilateral notching; , BMI, mean PI and bilateral notching.(b) BMI combined with mean resistance index (RI) and bilateral notching. , BMI; , mean RI and bilateral notching; , BMI,mean RI and bilateral notching.

0.78 (95% CI, 0.610.96), P=0.39) in one dataset from193 women.

Figure 4 shows two calibration plots. The model withBMI, mean RI and bilateral notching in Figure 4a hasgood calibration; women in the two deciles with highestcalculated probabilities of pre-eclampsia can be easilydistinguished from women with lower probabilities. Themodel with systolic blood pressure, mean PI and bilateralnotching in Figure 4b shows the same trends, but CIs arelarger because of a much smaller sample size.

Pre-eclampsia requiring delivery before 34 weeks

Because not all studies had recorded the variables pre-eclampsia requiring delivery before 34 weeks gestation

or gestational age at delivery, only the added value of

mean RI and bilateral notching or mean PI and bilateral

notching to BMI could be investigated. In two datasets

combined, containing data from 4442 women, of whom

32 developed pre-eclampsia requiring delivery before 34

weeks, discrimination with a model containing only BMI

was significantly improved by the addition of mean RI and

bilateral notching; the AUC increased from 0.66 (95% CI,

0.570.76) to 0.92 (95% CI, 0.870.98) (P


8/11


1.000.00

0.10

0.20

0.30

0.40

0.50

Sensitivity 0.60

0.70

0.80

0.90

1.00(a)

0.90 0.80 0.70 0.60 0.50

Specificity

0.40 0.30 0.20 0.10 0.00 1.000.00

0.10

0.20

0.30

0.40

0.50

Sensitivity 0.60

0.70

0.80

0.90

1.00(b)

0.90 0.80 0.70 0.60 0.50

Specificity

0.40 0.30 0.20 0.10 0.00

Figure 3 Receiveroperating characteristics curves showing discriminative abilities of systolic blood pressure (SBP), body mass index (BMI),uterine artery Doppler ultrasound measurements and their combinations in the identification of women at risk for pre-eclampsia. (a) SBP

and BMI combined with mean pulsatility index (PI) and bilateral notching. , SBP and BMI; , mean PI and bilateral notching;, SBP, BMI, mean PI and bilateral notching. (b) SBP and BMI combined with mean resistance index (RI) and bilateral notching. ,

SBP and BMI; , mean RI and bilateral notching; , SBP, BMI, mean RI and bilateral notching.

0.000.00

0.05

0.10Ob

servedrate 0.15

0.20

0.25(a)

0.05 0.10 0.15 0.20 0.25

Calculated probability

0.000.00

0.10

0.05

0.15

0.20

0.25

Ob

servedrate

0.30

0.35

0.45

0.40

0.50

0.55(b)

0.150.100.05 0.20 0.350.300.25 0.450.40 0.550.50

Calculated probability

Figure 4 Calibration plots of prediction models based on: (a) body mass index, mean uterine artery resistance index and bilateral notchingand (b) systolic blood pressure, mean uterine artery pulsatility index and bilateral notching.

0.62 (95% CI, 0.48 0.75) to 0.95 (95% CI, 0.92 0.98)(Pp75p80, >p80p85, >p85p90, >p90p95and>p95) and all other women with probabilitiesofp75, with probabilities calculated with the two mod-els described earlier: the model with BMI, mean RI andbilateral notching (4533 women) and the model with

blood pressure, mean PI and bilateral notching (214women). Figure 5 shows that women ranked in the high-est centiles of predicted probabilities from both modelsdelivered earlier than women with lower probabilities ofpre-eclampsia.

DISCUSSION

This paper describes the added value of uterine arteryDoppler ultrasound to patient characteristics in theidentification of nulliparous women at risk for pre-eclampsia. We found that the combination of Dopplerultrasound parameters PI or RI and bilateral notchingsignificantly improves the prediction of pre-eclampsia



9/11


28

0.0

0.2

0.4

Cumulativeproportionofwomennotdelivered

0.6

0.8

1.0(a)

30 32 34 36 38 40 42Gestational age at delivery (weeks)

28

0.0

0.2

0.4

Cumulativeproportionofwomennotdelivered

0.6

0.8

1.0(b)

30 32 34 36 38 40 42Gestational age at delivery (weeks)

Figure 5 Survival curves showing time to delivery in women grouped by percentiles (p) of predicted risk for pre-eclampsia based on (a) bodymass index, mean uterine artery resistance index and bilateral notching and (b) systolic blood pressure, mean uterine artery pulsatility indexand bilateral notching. , p75; , >p75 p80; , >p80 p85; , >p85 p90; , >p90 p95; , >p95.

based on the patient characteristics BMI and systolicblood pressure, alone or combined. The calibration plotsshow that women at the highest risk for pre-eclampsiacan be easily differentiated from women with lowerrisk predictions. This is supported by the time-to-eventanalyses, in which women with the highest calculated

risk of pre-eclampsia were shown to deliver earlierthan women at lower risk. This difference in time todelivery could be caused by the higher incidence of pre-eclampsia, fetal growth restriction or intrauterine fetaldeath and induced delivery in the group with higherpredicted probabilities of pre-eclampsia27,28. However, itis important that our results be externally validated inanother large dataset.

The methodology of IPD meta-analysis has severaladvantages over conventional meta-analysis. IPD meta-analysis can use all available data from a study, includinginformation that was not published, and is more flexible

in combining data into a uniform format. In addition, theuse of original data and contact with authors allow forsuperior quality checks and better interpretation of theresults29.

We approached over 95% of the authors of eligiblestudies, as identified in an extensive search, forparticipation in this project. The fact that only a minorityof eligible datasets were available for data sharing isa potential drawback of this project, but an interestingfinding as well. Worthy of remark here is our experiencethat for two studies the institutional review board ofthe center invited to participate did not agree withanonymized data sharing. Moreover, data from cohortstudies are apparently not stored as well as are datafrom randomized trials, although the advantagesthe

possibilities for answering new questions with the useof available good-quality data rather than having toinclude women in a new studyare clear and have beenhighlighted before30,31.

Within the shared datasets, we could only use theavailable patient characteristics. As all studies were

primarily designed to investigate uterine artery Doppler,the number and type of patient characteristics differedbetween studies and hampered our ability to study allpossible combinations in all available datasets. Althoughwe could only investigate a limited number of patientcharacteristics, we believe that those available include themost relevant factors for the prediction of pre-eclampsiathus far identified.

Out of all shared datasets with second-trimesterDoppler findings we chose a common and preferablyunselected group of women, as test performance is bestinvestigated in a population with an average risk for the

chosen outcome, a population that resembles the onethat will be tested once the test is introduced into dailypractice. As most Doppler studies identified in our searchhad been performed in women selected in various waysfor their risk of pre-eclampsia (for example based onpre-eclampsia in a previous pregnancy or pre-existentmedical conditions), we chose the relatively unselectedgroup of nulliparous women. The pre-eclampsia riskwas similar for all included studies, as reflected in thecomparable incidences of pre-eclampsia in thestudies. Thedata shared by Diab and Macleod consisted of womenselected on the basis of abnormal Doppler results outof all the women included in their studies. Assumingthat the higher incidence of pre-eclampsia was relatedto the abnormal Doppler findings but not to the patient



10/11


characteristics (which were comparable with those in theother included studies), these women were only includedin the estimation of predictive performance of patientcharacteristics alone. One could speculate that since thedata shared by Thilaganathan formed the largest part ofour dataset, our results may partly reflect the associationspresent in their data. However, in a random effects

model the weighted larger influence of larger studies isunweighted by the extent of variability of the effect sizesover all included studies32. Moreover, because not allvariables were present in the dataset of Thilaganathan, itwas not part of all prediction models.

The SCOPE study investigated the predictive perfor-mance of patient characteristics and Doppler ultrasoundin a large prospective cohort of nulliparous women witha similar 5% incidence of pre-eclampsia33. In that studythe AUC of the ROC curve of the best prediction modelwas around 0.71 on internal validation, but Dopplerultrasound did not add to the predictive performance of

patient characteristics (the AUC remained 0.71). In addi-tion to the patient characteristics blood pressure and BMIthat were included in our prediction models, the SCOPEmodel also contained information on age, family history,smoking and alcohol use, fruit intake, bleeding, previ-ous miscarriages, duration of sexual relationship, time toconception and the womans own birth weight. It seemsthat Doppler ultrasound either does not contain predic-tive information beyond that provided by these patientcharacteristics, or that abnormal Doppler findings are asign of impaired placentation and, as such, early diseasethat can be predicted by patient characteristics34,35. Incontrast, we did not find any improvement in our models

by the addition of data on smoking and/or age, and wehave shown that, in the absence of more predictive patientinformation, Doppler ultrasonography does have addedvalue in identifying women at risk.

So should we use this test in daily practice? OfferingDoppler ultrasonography (alone or in combination withother factors in a prediction model) as a screeningtest for pre-eclampsia would be justified if there werea safe, effective and affordable intervention to preventpre-eclampsia in high-risk women. So far, there is notenough evidence that giving calcium or vitamins to high-risk women (or all women) can prevent this disease well

enough to be cost-effective, neither has the benefit ofintensive monitoring been proven36,37. Although aspirintreatment is thought to have a relative risk of 0.90 foralmost all subgroups, there are indications that startingin early pregnancy is more effective than starting later inpregnancy4. This raises questions regarding the potentialeffectiveness of a screen-and-treat strategy in the secondtrimester.

Doppler measurements, as incorporated in a multi-variable prediction model, could also be suggested as alow-risk identification tool. Our calibration plots showthat women at higher risk (>1015%) can be well differ-entiated from women with lower risks (2%). Whereasusing the model as a screening test with a positivity thresh-old of 10 or 15% risk would imply that 8590% of

screen-positive women would be worried, monitored oreven treated unnecessarily, using Doppler as an exclu-sion test only failed to identify 2% of women whodeveloped pre-eclampsia. It is possible that the major-ity of the low-risk women among these 2% have mild,late-onset pre-eclampsia. Until further research showsmore effective screening or therapeutic options for pre-

eclampsia, Doppler ultrasound could be used to supportclinicians judgment of low-risk pregnancy.

ACKNOWLEDGMENTS

Emily Kleinrouweler is supported by a PhD Scholarshipfrom the AMC Graduate School. The AMC GraduateSchool had no involvement in the design and conductof the study; collection, management, analysis andinterpretation of the data; in the preparation, review andapproval of the manuscript or in the decision to submitthe manuscript for publication.

REFERENCES

1. Sibai B, Dekker G, Kupferminc M. Pre-eclampsia. Lancet2005;365: 785799.

2. Askie LM, Duley L, Henderson-Smart DJ, Stewart LA.Antiplatelet agents for prevention of pre-eclampsia: a meta-analysis of individual patient data. Lancet 2007; 369:17911798.

3. Duley L, Henderson-Smart DJ, Meher S, King JF. Antiplateletagents for preventing pre-eclampsia and its complications.Cochrane Database Syst Rev 2007: CD004659.

4. Bujold E, Roberge S, Lacasse Y, Bureau M, Audibert F,Marcoux S, Forest JC, Gigu `ere Y. Prevention of preeclampsiaand intrauterine growth restriction with aspirin started inearly pregnancy: a meta-analysis. Obstet Gynecol 2010; 116:402414.

5. Cnossen JS, ter Riet G, Mol BW, van der Post JA, LeeflangMM, Meads CA, Hyde C, Khan KS. Are tests for predictingpre-eclampsia good enough to make screening viable? A reviewof reviews and critical appraisal. Acta Obstet Gynecol Scand2009; 88: 758765.

6. Conde-Agudelo A, Villar J, Lindheimer M. World Health Orga-nization systematic review of screening tests for preeclampsia.Obstet Gynecol2004; 104: 13671391.

7. Audibert F, Boucoiran I, An N, Aleksandrov N, DelvinE, BujoldE, Rey E. Screening for preeclampsia using first-trimester serummarkers and uterine artery Doppler in nulliparous women. Am

J Obstet Gynecol2010; 203: 383.e18.8. Poon LC, Karagiannis G, Leal A, Romero XC, NicolaidesKH. Hypertensive disorders in pregnancy: screening by uterineartery Doppler imaging and blood pressure at 1113 weeks.Ultrasound Obstet Gynecol2009; 34: 497502.

9. Thilaganathan B, Wormald B, Zanardini C, Sheldon J, RalphE, Papageorghiou AT. Early-pregnancy multiple serum markersand second-trimester uterine artery Doppler in predictingpreeclampsia. Obstet Gynecol2010; 115: 12331238.

10. Tuuli MG, Odibo AO. First- and second-trimester screeningfor preeclampsia and intrauterine growth restriction. Clin LabMed2010; 30: 727746.

11. Zhong Y, Tuuli M, Odibo AO. First-trimester assessmentof placenta function and the prediction of preeclampsiaand intrauterine growth restriction. Prenat Diagn 2010; 30:

293308.12. Broeze KA, Opmeer BC, van der Veen F, Bossuyt PM,Bhattacharya S, Mol BW. Individual patient data meta-analysis:



11/11


a promising approach for evidence synthesis in reproductivemedicine. Hum Reprod Update 2010; 16: 561567.

13. Stroup DF, Berlin JA, Morton SC, Olkin I, Williamson GD,Rennie D, Moher D, Becker BJ, Sipe TA, Thacker SB. Meta-analysis of observational studies in epidemiology: a proposalfor reporting. Meta-analysis Of Observational Studies inEpidemiology (MOOSE) group.JAMA 2000; 283: 2008 2012.

14. Brown MA, Lindheimer MD, de Swiet M, Van AsscheA, Moutquin JM. The classification and diagnosis of thehypertensive disorders of pregnancy: statement from theInternational Society for the Study of Hypertension inPregnancy (ISSHP). Hypertens Pregnancy 2001; 20: IX XIV.

15. Whiting PF, Weswood ME, Rutjes AW, Reitsma JB, BossuytPN, Kleijnen J. Evaluation of QUADAS, a tool for the qualityassessment of diagnostic accuracy studies. BMC Med ResMethodol2006; 6: 9.

16. Steyerberg EW. Clinical Prediction Models. A practicalapproach to development, validation and updating. Springer:New York, 2009.

17. Harrell FE Jr. Regression Modeling Strategies. Springer-Verlag:New York, 2001.

18. Twisk JWR. Software for multilevel analysis. In AppliedMultilevel Analysis (Twisk JWR). Cambridge University Press :

Cambridge, 2007; 146150.19. Arenas J, Fernandez-Inarrea J, Rodriguez-Mon C, Dupla B,Diez E, Gonzalez-Garcia A. Doppler screening of the uterinearteries to predict complications during pregnancy. [Spanish].Clinica e Investigacion en Ginecologia y Obstetricia 2003; 30:178184.

20. Deurloo KL, Spreeuwenberg MD, Bolte AC, van Vugt JM.Color Doppler ultrasound of spiral artery blood flow forprediction of hypertensive disorders and intra uterine growthrestriction: a longitudinal study. Prenat Diagn 2007; 27:10111016.

21. Diab AE, El-Behery MM, Ebrahiem MA, Shehata AE.Angiogenic factors for theprediction of pre-eclampsia in womenwithabnormal midtrimester uterine arteryDoppler velocimetry.Int J Gynaecol Obstet2008; 102: 146151.

22. Khan F, Belch JJ, MacLeod M, Mires G. Changes in endothelialfunction precede the clinical disease in women in whompreeclampsia develops. Hypertension 2005; 46: 11231128.

23. Ohkuchi A, Minakami H, Sato I, Mori H, NakanoT, Tateno M.Predicting the risk of pre-eclampsia and a small-for-gestational-age infant by quantitative assessment of the diastolic notch inuterine artery flow velocity waveforms in unselected women.Ultrasound Obstet Gynecol2000; 16: 171178.

24. Napolitano R, Santo S, DSouza R, Bhide A, Thilaganathan B.Sensitivity of higher, lower and mean second-trimester uterineartery Doppler resistance indices in screening for pre-eclampsia.Ultrasound Obstet Gynecol2010; 36: 573576.

25. Vollebregt KC, Gisolf J, Guelen I, Boer K, van Montfrans G,Wolf H. Limited accuracy of the hyperbaric index, ambulatoryblood pressure and sphygmomanometry measurements in pre-dicting gestational hypertension and preeclampsia. J Hypertens2010; 28: 127134.

26. Rang S, van Montfrans GA, Wolf H. Serial hemodynamicmeasurement in normalpregnancy,preeclampsia, and intrauter-ine growth restriction. Am J Obstet Gynecol 2008; 198:519.e19.

27. Audibert F, Benchimol Y, Benattar C, Champagne C, FrydmanR. Prediction of preeclampsia or intrauterine growth restrictionby second trimester serum screening and uterine Dopplervelocimetry. Fetal Diagn Ther 2005; 20: 4853.

28. Singh T, Leslie K, Bhide A, DAntonio F, Thilaganathan B. Roleof second-trimester uterine artery Doppler in assessing stillbirthrisk. Obstet Gynecol2012; 119: 256261.

29. Khan KS, Bachmann LM, ter Riet G. Systematic reviews withindividual patient data meta-analysis to evaluate diagnostictests. Eur J Obstet Gynecol Reprod Biol 2003; 108:121125.

30. Broeze KA,OpmeerBC,Bachmann LM, BroekmansFJ, BossuytPM, Coppus SF, Johnson NP, Khan KS, ter Riet G, vander VeenF, van Wely M, Mol BW. Individual patient data meta-analysisof diagnostic and prognostic studies in obstetrics, gynaecologyand reproductive medicine. BMC Med Res Methodol 2009;9: 22.

31. Individual Patient Data meta-analysis of diagnostic andprognostic studies in obstetrics, gynaecology and reproductivemedicine. http://www.studies-obsgyn.nl/IPD [Accessed 10 May2012].

32. Senn S. Trying to be precise about vagueness. Stat Med 2007;26: 14171430.

33. North RA, McCowan LM, Dekker GA, Poston L, Chan EH,StewartAW, Black MA, Taylor RS, Walker JJ, Baker PN, KennyLC. Clinical risk prediction for pre-eclampsia in nulliparouswomen: development of model in international prospectivecohort. BMJ2011; 342: d1875.

34. Steegers EA, von Dadelszen P, Duvekot JJ, Pijnenborg R. Pre-eclampsia. Lancet2010; 376: 631644.

35. James JL, Whitley GS, Cartwright JE. Pre-eclampsia: fittingtogether the placental, immune and cardiovascular pieces. JPathol2010; 221: 363378.

36. Meads CA, Cnossen JS, Meher S, Juarez-Garcia A, ter RietG, Duley L, Roberts TE, Mol BW, van der Post JA, LeeflangMM, Barton PM, Hyde CJ, Gupta JK, Khan KS. Methodsof prediction and prevention of pre-eclampsia: systematicreviews of accuracy and effectiveness literature with economicmodelling. Health Technol Assess 2008; 12: iiiiv, 1 270.

37. Shmueli A, Meiri H, Gonen R. Economic assessment ofscreening for pre-eclampsia. Prenat Diagn 2012; 32: 2938.

S U P P O R T I N G I N F O R M A T I O N O N T H E I N T E R N E TThe following supporting information may be found in the online version of this article:

Appendix S1 This appendix contains additional information on the search strategy, quality assessment andanalysis methods that are described concisely in the paper. In addition, the investigation of the association ofthe available patient characteristics with pre-eclampsia and the selection of the most predictive patientcharacteristics are described.


Documents

Uterine Artery Doppler ISUOG 2013