Professional Documents
Culture Documents
Second-Trimester Uterine Artery Doppler Screening in
Second-Trimester Uterine Artery Doppler Screening in
Objective: Doppler ultrasound provides a non-invasive method for the study of the uteroplacental
circulation. In normal pregnancy, impedance to flow in the uterine arteries decreases with gestation,
which may be the consequence of trophoblastic invasion of the spiral arteries and their conversion
J Matern Fetal Neonatal Med Downloaded from informahealthcare.com by Nyu Medical Center on 05/29/15
into low-resistance vessels. Pre-eclampsia and fetal growth restriction are associated with failure of
trophoblastic invasion of spiral arteries, and Doppler studies, in these conditions, have shown that
impedance to flow in the uterine arteries is increased. A series of screening studies involving assess-
ment of impedance to flow in the uterine arteries have examined the potential value of Doppler in
identifying pregnancies at risk of the complications of impaired placentation. This review examines
the findings of Doppler studies in unselected populations.
Methods: Searches of a computerized medical database were performed to identify relevant studies.
Only those studies that provided sufficient data to allow calculation of the performance of the test
were included in the analysis. Likelihood ratios were calculated for each study and are reported
for pre-eclampsia, fetal growth restriction and perinatal death as well as for more severe forms of
pre-eclampsia and fetal growth restriction.
Results: The literature search identified 19 relevant studies, four of which were excluded from the
For personal use only.
further analysis. The main characteristics and results of the 15 remaining studies provided discrep-
ant results, which may be the consequence of differences in Doppler technique for sampling, the
definition of abnormal flow velocity waveform, differences in the populations examined, the gesta-
tional age at which women were studied and different criteria for the diagnosis of pre-eclampsia and
fetal growth restriction. Nevertheless, the studies provided evidence that increased impedance to
flow in the uterine arteries is associated with increased risk for subsequent development of
pre-eclampsia, fetal growth restriction and perinatal death. In addition, women with normal imped-
ance to flow in the uterine arteries constituted a group that have a low risk of developing obstetric
complications related to uteroplacental insufficiency.
Conclusions: The review suggests that increased impedance to flow in the uterine arteries in
pregnancies attending for routine antenatal care identifies about 40% of those who subsequently
develop pre-eclampsia and about 20% of those who develop fetal growth restriction. Following
a positive test, the likelihood of these complications is increased by about 6 and 3.5 times,
respectively.
INTRODUCTION
In normal pregnancy, the spiral arteries, destined to arteries into vessels of greater diameter with low resistance
become the uteroplacental arteries in the placental bed, and high compliance, that are unresponsive to maternal
undergo a complex series of morphological changes. The vasomotor activity1,5. This vascular transformation in the
vessels are invaded by trophoblast, which becomes incorpo- uterus is necessary to ensure a dramatic increase in blood
rated into the vessel wall and replaces the endothelium, supply to the intervillous space.
muscular layer and neural tissue1–4. The result of these In pre-eclampsia and fetal growth restriction, there is
physiological changes is the conversion of the small spiral failure of the perivascular and endovascular trophoblastic
Correspondence: Pr ofessor K. H. Nicolaides, Har ris Birthright Research Centre for Fetal Medicine, King’s College Hospital Medical
School, Denm ark Hill, L ondon SE5 9RS, UK
R EVIEW 78 Received 22–05–02 Accepted 01–06–02
Second-trimester uterine artery Doppler screening Papageorghiou et al.
invasion of spiral arteries. Although, in 50–70% of the review examines the findings of all available Doppler
spiral arteries, there is trophoblastic invasion, this is studies of the uterine arteries during the second trimester
confined to the decidual part of the vessels and does not in unselected populations.
extend into the myometrial segments2,3,6,7. In addition,
many vessels are occluded by fibrinoid material and exhibit
adjacent foam cell accumulation and a perivascular mono- METHODS
nuclear cell infiltrate, which is referred to as atherosis8–10.
Searches of a computerized medical database (PubMed)
Doppler ultrasound provides a non-invasive method
were conducted independently by two of the researchers
for the study of the uteroplacental circulation. In normal
(A.P., C.Y.). In addition to this, references from these
pregnancy, impedance to flow in the uterine arteries
articles that were not identified by the original search were
decreases with gestation and this presumably reflects the
included in the list of articles for review. The criteria for
trophoblastic invasion of the spiral arteries and their con-
reviewing the full articles were: pregnant subjects from
version into low-resistance vessels11. Cross-sectional
unselected populations, uterine artery Doppler assessment,
studies in pregnancies with pre-eclampsia or fetal growth
J Matern Fetal Neonatal Med Downloaded from informahealthcare.com by Nyu Medical Center on 05/29/15
and Nam12.
RESULTS
The literature search identified 19 relevant studies. From
these, four studies were excluded from the further analysis:
the studies of Campbell and colleagues13 and Hanretty and
associates14 because these examined arcuate rather than
uterine arteries; the study by Schulman and co-workers15
because the gestational age at examination was ill-defined;
and the study by Liberati and associates16 because, although
initial inclusion criteria were on an unselected basis, the
Figure 1 Normal flow velocity waveform from the uterine artery aim of the study was to evaluate the screening indices only
at 24 weeks of gestation in those women with lateral placentas, and subsequent
analysis included only such women.
The main characteristics and results of the 15 remaining
studies are summarized in Tables 1–617–31. Two studies20,21
reported on the same study population using different
methodology and Doppler measurements (one-stage versus
two-stage screening). Only one of these 20 has been
included in the calculation of pooled likelihood ratios
for the three outcomes of interest, which are shown in
Tables 7–9.
The main features of the screening studies, including
Figure 2 Flow velocity waveform from the uterine artery at 24 number of patients, gestational age, Doppler technique
weeks of gestation in a pregnancy with impaired placentation. In and definition of an abnormal result, are summarized in
early diastole, there is a notch and, in late diastole, there is Table 1. The early studies were limited by the use of con-
decreased flow tinuous wave Doppler, which is a blind investigation.
7
J ournal of Maternal–Fetal and Neonatal Medicin e
Second-trimester uterine artery Doppler screening Papageorghiou et al.
Table 1 Characteristics of uterine artery Doppler screening studies used in this review
Gestation Screen Outcome
Reference n Doppler Abnormal result (weeks) positive (%) measure
199322
North et al., 199423 446 (457)‡ color-PW RI > 0.57 19–24 11.9 (12.3)‡ PET, FGR
Todros et al., 199524 916 CW or PW S/D ratio > 2.7 19–24 6.4 FGR
Harrington et al., 1204 CW and color-PW RI > 95th centile 20 and 24 9.1 PET, FGR, PD
199625 or notches
Frusca et al., 199726 419 CW and color-PW mean RI > 0.58 20 and 24 8.6 PET, FGR
Irion et al., 199827 1159 color-PW mean RI > 0.57 26 12.8 PET, FGR
Kurdi et al., 199828 946 color-PW mean RI > 0.55 and 19–21 12.4† PET, FGR
bilateral notches†
Albaiges et al., 1757 color-PW mean PI > 95th 22–24 7.3 PET, FGR, PD
For personal use only.
200029 centile or
bilateral notches
Aquilina et al., 640 color-PW mean RI ³ 0.65 and 19–20 9.8 PET, FGR, PD
200130 notches
Papageorghiou 7851 TV, color-PW mean PI > 95th 22–24 5.1 PET, FGR
et al., 200131 centile
n, number of cases used in the analysis after all exclusions; CW, continuous-wave Doppler without visualizing the vessel;
PW, pulsed-wave Doppler after identification of the vessel using B-mode ultrasound; color-PW, color flow imaging to identify the vessel
followed by pulsed-wave Doppler; TV, transvaginal; RI, resistance index; S/D systolic to diastolic ratio; PI, pulsatility index;
PET, pre-eclampsia; FGR, fetal growth restriction; PD, perinatal death
*These two studies report on the same population using one-stage and two-stage Doppler screening
†
In Kurdi et al., 199828, this definition of an abnormal result was used for the prediction of pre-eclampsia and fetal growth restriction
below the 5th centile. In the prediction of fetal growth restriction below the 10th centile, an abnormal result was defined as the presence
of bilateral notches and mean RI > 0.55, or unilateral notch and mean RI > 0.65, or mean RI > 0.7 (screen positive rate 22.8%)
‡
Pre-eclampsia (fetal growth restriction)
Subsequently, real-time ultrasound was used to identify the of abnormal flow velocity waveform (resistance or
vessels and pulsed wave Doppler was applied to obtain pulsatility index above a certain cut-off or a centile or the
the necessary waveforms. More recently, identification of presence of an early diastolic notch). In addition, there
the vessels has been made easier with the use of color flow were differences in the populations examined (for example,
mapping. This can be achieved either transabdominally the prevalence of pre-eclampsia varied from 1.4% to 5.5%),
(at the apparent cross-over with the external iliac artery, the gestational age at which women were studied (16–24
Figure 3) or transvaginally (lateral to the uterine cervix at weeks) and in the criteria for the diagnosis of pre-eclampsia
the level of the internal cervical os, Figure 4). and fetal growth restriction. Four of the studies reported on
Discrepant results between the studies may be the conse- two-stage screening; initial screening was carried out by
quence of differences in Doppler technique for sampling continuous wave Doppler at 18–20 weeks of gestation and,
(for example, continuous wave, pulsed wave or color in those with increased impedance to flow, the Doppler
Doppler to insonate the uterine arteries) and the definition studies were repeated at 24 weeks19,21,25,26.
80
J ournal of Maternal–Fetal and Neonatal Medicin e
Second-trimester uterine artery Doppler screening Papageorghiou et al.
Table 2 Results of uterine artery Doppler screening studies for the prediction of pre-eclampsia, providing data on the definition of
pre-eclampsia used, screen-positive rate, prevalence, sensitivity, specificity, positive predictive value (PPV) and negative predictive
value (NPV)
Screen-positive Prevalence Sensitivity Specificity PPV NPV
Reference rate (%) (%) (%) (%) (%) (%)
Table 3 Likelihood ratios (LR) of uteroplacental Doppler screening in the prediction of pre-eclampsia
For personal use only.
Summary of the screening studies Bewley and co-workers18 calculated the average resis-
tance index from the left and right uterine and arcuate
Newnham and associates17 measured the systolic/diastolic
arteries in 925 pregnancies at 16–24 weeks’ gestation. The
ratio in 501 women using continuous wave Doppler at 18
analysis for pre-eclampsia was based on 917 pregnancies
weeks of gestation. A screen-positive result (defined as a
and that for fetal growth restriction on 913. When the
systolic/diastolic ratio above the 95th centile) was found in
resistance index was greater than the 95th centile, there
4.8% of women and the sensitivity for fetal growth restric-
was a 10-fold increase in risk for a severe adverse outcome,
tion less than the 10th centile for gestation was 6.1% for a
defined by fetal death, placental abruption, fetal growth
specificity of 95%.
restriction or pre-eclampsia (prevalence 7%, sensitivity
8
J ournal of Maternal–Fetal and Neonatal Medicin e
Second-trimester uterine artery Doppler screening Papageorghiou et al.
Table 4 Results of uterine artery Doppler screening studies for the prediction of more severe forms of pre-eclampsia (PET), providing
data on the definition of screen-positive rate, severity of disease and likelihood ratios (LR) of a positive and negative test with 95% confi-
dence intervals (CI)
Screen- Incidence Positive test Negative test
positive rate Sensitivity
Reference (%) Disease n % (%) LR 95% CI LR 95% CI
PET/gestational hypertension
Frusca et al., 199726 8.6 PET 8 1.9 50 6.42 2.62–11.71 0.54 0.23–0.85
GH 24 5.7 25 3.29 1.48–6.59 0.81 0.60–0.96
Bewley et al., 199118 5.6 PET 42 4.6 24 5.08 2.69–9.03 0.80 0.64–0.91
GH 50 5.5 10 1.88 0.79–4.25 0.95 0.83–1.01
Steel et al., 199019 11.6 PET 19 1.9 63 5.93 3.73–8.17 0.41 0.21–0.66
GH 74 7.3 39 4.14 2.89–5.75 0.67 0.55–0.79
J Matern Fetal Neonatal Med Downloaded from informahealthcare.com by Nyu Medical Center on 05/29/15
PET/PET preterm
Harrington et al., 3.9 PET 44 3.7 55 26.36 16.15–41.97 0.46 0.32–0.61
199625* PET del. < 35 weeks 16 1.3 81 27.58 17.39–40.14 0.19 0.07–0.44
Kurdi et al., 199828 12.4 PET 13 1.4 62 5.51 3.52–7.59 0.43 0.23–0.67
PET del. < 37 weeks 7 0.8 88 7.46 4.41–9.41 0.14 0.03–0.53
Albaiges et al., 7.3 PET 65 3.7 45 7.63 5.38–10.41 0.59 0.46–0.71
200029 PET del. < 34 weeks 10 0.6 90 13.21 8.53–16.37 0.11 0.02–0.43
Papageorghiou et al., 5.1 PET 113 1.4 41 8.87 6.86–11.17 0.62 0.52–0.71
200131 PET del. < 38 weeks 72 0.9 54 11.64 9.01–14.41 0.48 0.37–0.60
For personal use only.
*In Harrington et al., 199625, the definition of an abnormal result for pre-eclampsia requiring preterm delivery was that of bilateral
uterine artery notches rather than the definition given in Table 1. GH, gestational hypertension; FGR, fetal growth restriction;
del., delivery
21%, specificity 95%, positive predictive value 25%). How- Bower and co-workers20 examined the uterine arteries in
ever, the sensitivities of the test for pre-eclampsia or 2058 pregnancies at 18–22 weeks. An abnormal result,
fetal growth restriction (below the 5th centile) were only defined by a resistance index above the 95th centile or the
24% and 19%, respectively, with a specificity of about 95% presence of an early diastolic notch in either of the two
for both. uterine arteries, was found in 16% of the pregnancies. The
Steel and colleagues19 examined the uterine arteries in sensitivity of the test was 75% for pre-eclampsia with speci-
1014 nulliparous women by continuous wave Doppler at 18 ficity of 86%. This study highlighted that abnormal
weeks of gestation, and, in those with increased impedance Doppler results provide good prediction of the more severe
(resistance index greater than 0.58), the Doppler studies types of pregnancy complications. Thus, the sensitivity for
were repeated at 24 weeks. A screen-positive result mild pre-eclampsia was only 29%, but, for moderate/severe
(increased impedance at 24 weeks) was found in 12% of disease, the sensitivity was 82%. Similarly, the sensitivity
cases and the sensitivity of the test for pre-eclampsia was for birth weight below the 10th centile was 37% and, for
63%, and, for fetal growth restriction, it was 43%. There birth weight below the 5th centile, it was 46%. There were
were 15 perinatal deaths (prevalence 1.5%), and abnormal 27 fetal deaths (prevalence of 1.3%) and an abnormal
Doppler results identified 33% of these with a specificity Doppler result predicted 18.5% of the cases with a specifi-
of 89%. city of 84%.
82
J ournal of Maternal–Fetal and Neonatal Medicin e
Second-trimester uterine artery Doppler screening Papageorghiou et al.
Table 5 Results of uterine artery Doppler screening studies for the prediction of fetal growth restriction providing data on the defini-
tion of fetal growth restriction (FGR) used, screen-positive rate, prevalence, sensitivity, specificity, positive predictive value (PPV) and
negative predictive value (NPV)
Screen-positive Prevalence Sensitivity Specificity PPV NPV
Reference FGR definition rate (%) (%) (%) (%) (%) (%)
*In Kurdi et al.28, different definitions of an abnormal result were used for the prediction of fetal growth restriction below the 5th and
below the 10th centile, respectively (see Table 1)
Table 6 Results of uterine artery Doppler screening studies for the prediction of fetal or perinatal death providing data on screen
positive rate, prevalence, sensitivity, specificity, positive predictive value (PPV) and negative predictive value (NPV)
Outcome Screen-positive Prevalence Sensitivity Specificity PPV NPV
Reference measure rate (%) (%) (%) (%) (%) (%)
In another study by Bower and associates21, in a patients (compared to 16% at 20 weeks). It was reported
subgroup of the 2058 unselected women who were that increased impedance provides good prediction of
initially examined by continuous wave Doppler at 20 weeks pre-eclampsia (but not non-proteinuric pregnancy-induced
of gestation (see above) and found to have increased hypertension).
impedance to flow (resistance index greater than the Valensise and colleagues22 examined the uterine arteries
95th centile or early diastolic notch in either of the two in 272 primigravidas at 22 weeks of gestation. An abnormal
uterine arteries), the Doppler studies were repeated by result, defined by increased impedance (mean resistance
color Doppler at 24 weeks. As 32 women did not have index of more than 0.58) was found in 9.6% of patients.
the second stage of screening, the analysis is based on the The sensitivity of the test in predicting pre-eclampsia
remaining 2026 women (see Table 1). Persistently was 89% and, for fetal growth restriction, it was 67%;
increased impedance was observed in 5.1% of the the specificities were 93% and 95%, respectively. The
8
J ournal of Maternal–Fetal and Neonatal Medicin e
Second-trimester uterine artery Doppler screening Papageorghiou et al.
Table 7 Likelihood ratios (LR) of uteroplacental Doppler screening in the prediction of fetal growth restriction (FGR)
Positive test Negative test
Table 8 Likelihood ratios (LR) of uteroplacental Doppler screening in the prediction of perinatal death
Positive test Negative test
sensitivity for predicting non-proteinuric pregnancy- they found increased impedance (resistance index
induced hypertension was 50%. greater than 0.57 on the placental side) in 12% of cases.
North and co-workers23 examined the uterine arteries at The sensitivity of the test for pre-eclampsia was 27% and,
19–24 weeks of gestation in 457 nulliparous women and for fetal growth restriction, it was 50%; the respective
84
J ournal of Maternal–Fetal and Neonatal Medicin e
Second-trimester uterine artery Doppler screening Papageorghiou et al.
8
J ournal of Maternal–Fetal and Neonatal Medicin e
Second-trimester uterine artery Doppler screening Papageorghiou et al.
50%. Six fetal deaths (prevalence 0.34%) occurred and The studies that examined different severities of this
abnormal Doppler results identified 83.3% of these, with a complication reported that Doppler was better in predict-
specificity of 92.8%. The screening results were similar ing more severe disease (Table 4). Thus, Bower and
if the high-risk group was defined as either those with associates20 reported that increased impedance provides
increased pulsatility index or bilateral notches. good prediction of pre-eclampsia but not pregnancy-
Aquilina and colleagues30 examined 640 women at induced hypertension without proteinuria. The enhanced
19–21 weeks of gestation using color Doppler. Pre- predictive ability of uterine artery screening in more severe
eclampsia occurred in 5.5% of pregnancies. Abnormal disease is also shown in the study by Papageorghiou and
uterine artery blood flow, as defined by mean resistance co-workers31 where sensitivities for more severe disease, as
index greater than or equal to 0.65 with unilateral or defined by the presence of pre-eclampsia with fetal growth
bilateral notches, was present in 9.8% of cases, and the restriction, were reported. In this study, pre-eclampsia with
sensitivity of predicting pre-eclampsia was 60% with a fetal growth restriction occurred in 0.5% of cases, pre-
specificity of 93%. eclampsia without fetal growth restriction in 0.9% and fetal
Papageorghiou and associates31 examined 7851 single- growth restriction without pre-eclampsia in 8.9% of cases,
J Matern Fetal Neonatal Med Downloaded from informahealthcare.com by Nyu Medical Center on 05/29/15
ton pregnancies attending for routine antenatal ultrasound and the respective sensitivities were 69%, 24% and 13%.
at 23 weeks using a transvaginal color Doppler approach. Another measure of the severity of the disease is the
The mean uterine artery pulsatility index of the two gestation at which delivery is undertaken. Harrington and
arteries was calculated and the presence or absence of colleagues25 found that bilateral notching at 24 weeks
notches was documented. A mean pulsatility index above identified 55% of women who later developed pre-
the 95th centile (1.63) was found in 5.1% of women and eclampsia and 81% with pre-eclampsia requiring delivery
this cut-off identified 41% of women who subsequently before 35 weeks. Similarly, Kurdi and associates28 found
developed pre-eclampsia and 16% of those delivering an that bilateral notches and a mean resistance index greater
infant with a birth weight below the 10th centile with than 0.55 identified 62% of women who later developed
specificities of 95% and 96%, respectively. The sensitivity pre-eclampsia and 88% with pre-eclampsia requiring
for pre-eclampsia and fetal growth restriction increased delivery before 37 weeks. Albaiges and co-workers29
For personal use only.
with the severity of the disease. Thus, the sensitivity for showed that the sensitivity of increased pulsatility index or
pre-eclampsia requiring delivery before 38 weeks was 54% bilateral notches in predicting pre-eclampsia was 45%,
and the sensitivity increased to 70%, 81% and 90% for while, for pre-eclampsia requiring delivery before 34 weeks,
pre-eclampsia requiring delivery before 36, 34 and 32 the sensitivity was 90%. Papageorghiou and colleagues
weeks, respectively; the respective sensitivities for fetal reported that increased pulsatility index identified 41% of
growth restriction below the 10th centile were 35%, 53%, women who later developed pre-eclampsia and that the
64% and 74%. The test was better at predicting pre- sensitivities increased to 54%, 70%, 81% and 90% for
eclampsia accompanied by fetal growth restriction (69%) pre-eclampsia requiring delivery before 38, 36, 34 and 32
than pre-eclampsia in the absence of fetal growth restric- weeks, respectively.
tion (24%). Furthermore, the sensitivity for fetal growth
restriction defined by the 5th, rather than the 10th, centile
Prediction of fetal growth restriction
was higher (19% versus 16%). Bilateral notches were found
in 9.3% of women and, using this screening index, showed All studies defined this complication as the delivery of
similar sensitivities to using high pulsatility index, but with an infant with birth weight below the 10th centile for
a higher screen-positive rate. It was concluded that a gestation. Some studies also provided data on birth weight
one-stage Doppler screening program at 23 weeks identifies below the 5th and 3rd centiles. The range in the preva-
most women who subsequently develop severe pre- lence of fetal growth restriction, defined as birth weight
eclampsia and/or fetal growth restriction. below the 10th centile, was 4.6–16.5%, the sensitivity was
6–67% and the specificity was 82–96% (Table 5). In the
pooled data from all studies, the likelihood ratio for
Prediction of pre-eclampsia
the subsequent delivery of a growth-restricted infant in
The studies differed in the definition of hypertensive women with increased impedance to flow was about 3.7
disease, prevalence of this complication (1.4–5.5%), sensi- and, for those with normal Doppler results, the ratio was
tivity (24–89%) and specificity (86–96%) of the test about 0.8 (Table 7).
(Table 2). In the pooled data from all studies, the likeli- There was agreement between the studies that the sensi-
hood ratios for the subsequent development of pre- tivity of increased impedance was higher in the prediction
eclampsia in women with increased impedance to flow was of more severe degrees of growth restriction (Table 5). For
about 6 and, for those with normal Doppler results, the example, in the study of Bower and associates20, the sensi-
ratio was about 0.5 (Table 3). tivity for the prediction of birth weight below the 10th
86
J ournal of Maternal–Fetal and Neonatal Medicin e
Second-trimester uterine artery Doppler screening Papageorghiou et al.
centile was 37%, and this increased to 47% for birth weight During the last decade, uterine artery screening has
below the 3rd centile. The respective sensitivities in the evolved from a blind method (using continuous wave
study of Bewley18 were 15% and 20%; in the study of Doppler) to identification of the vessel using color flow
Albaiges and co-workers29, they were 22% and 30%. In the mapping. The use of continuous wave Doppler, without
study of Steel and colleagues19, they were 33% and 43% visualization of the uterine artery prior to insonation, may
and in the study of Frusca26 they were 43% and 55%. make it more difficult to distinguish uterine artery blood
The same point of improved sensitivity in the prediction flow from adjacent high resistance internal iliac vessels,
of severe disease is made by Papageorghiou and co- and the lower resistance arcuate arteries14.
workers31, who reported that mean pulsatility index above There has been a move away from two-stage to
the 95th centile of the normal range at 23 weeks identified one-stage screening. Historically, two-stage screening
16% of pregnancies delivering an infant with birth weight was performed because, in the mid-1990s, there was a
below the 10th centile and that this increased to 35%, sparsity of color Doppler equipment, and because the
53%, 64% and 74% for birth weight below the 10th centile routine anomaly scan was performed at 20 weeks (lending
delivering before 38, 36, 34 and 32 weeks, respectively. In itself to the performance of the first stage). However,
J Matern Fetal Neonatal Med Downloaded from informahealthcare.com by Nyu Medical Center on 05/29/15
the study by Albaiges and associates29, sensitivities for two-stage screening relies on those patients with an abnor-
mean pulsatility index above the 95th centile or bilateral mal first-stage test result returning for the second stage. In
notches were 22% for birth weight below the 10th centile three of the four two-stage screening studies, the drop-out
and 70% for birth weight below the 10th centile delivering from subsequent follow-up was between 10% and
before 34 weeks. Harrington and colleagues25 found that 27%21,25,26.
bilateral notching at 24 weeks identified 22% of women The gestational age at screening has moved to 23–24
who later delivered infants with a birth weight below the weeks, as earlier screening has been associated with a
10th centile, and the sensitivity improved to 58% for higher false-positive rate. The prevalence of bilateral
growth restriction requiring delivery before 35 weeks. notches at 18–22 weeks is more than 10%, and, at 23–24
Similarly, Kurdi and associates28 found that bilateral weeks, it is less than 5%. Two large recent studies have
notches and a mean resistance index greater than 0.55 at shown that, at 23–24 weeks, the sensitivities are similar if
For personal use only.
19–21 weeks identified 37% of women who later delivered high impedance is defined by the presence of bilateral
infants with a birth weight below the 5th centile, and the notches or a mean pulsatility index above the 95th centile
sensitivity improved to 88% for growth restriction requir- of the normal range for gestation29,31. Using the pulsatility
ing delivery before 37 weeks. index removes the element of subjectivity associated with
the interpretation of notches. Although, in both these
studies, inclusion of the presence of bilateral notches in
Prediction of perinatal death
the definition of an abnormal resulted in a small increase in
There are only four studies providing sufficient data to sensitivity, this was achieved at the expense of a substantial
examine the performance of Doppler screening in the pre- increase in the screen-positive rate.
diction of fetal or perinatal death (Table 6). There were The results obtained from one-stage screening tests at
large differences in the sensitivity (8–83%) of the test, 23–24 weeks suggest that increased impedance to flow in
which may well reflect the small number of cases in each the uterine arteries in pregnancies attending for routine
study. In the largest study, involving 27 fetal deaths, the antenatal care identifies about 40% of those who subse-
sensitivity and specificity of increased impedance were 19% quently develop pre-eclampsia and about 20% of those who
and 84%, respectively19. In the pooled data from the four develop fetal growth restriction. Following a positive test,
studies, the likelihood ratio for subsequent fetal or perinatal the likelihood of these complications is increased by about
death in women with increased impedance to flow was 2.4 6 and 3.5 times, respectively. Abnormal Doppler is better
and, for those with normal Doppler results, the ratio was in predicting severe rather than mild disease. The sensiti-
0.8 (Table 8). vity for severe disease requiring early delivery is about 80%
for pre-eclampsia and 60% for fetal growth restriction. The
sensitivity of screening for perinatal death is necessarily
CONCLUSIONS
based on small numbers, but the likelihood of this compli-
There is extensive evidence that increased impedance to cation in those with an abnormal Doppler result is about
flow in the uterine arteries is associated with increased risk 2.4 times higher than the background risk.
for subsequent development of pre-eclampsia, fetal growth
restriction and perinatal death. In addition, women with
ACKNOWLEDGEMENT
normal impedance to flow in the uterine arteries constitute
a group that has a low risk of developing obstetric compli- This study was supported by a grant from the Fetal
cations related to uteroplacental insufficiency. Medicine Foundation (Charity No. 1037116).
8
J ournal of Maternal–Fetal and Neonatal Medicin e
Second-trimester uterine artery Doppler screening Papageorghiou et al.
Immunology, 1st edn. New York: Cambridge University Press, 22. Valensise H, BezzeccheriV, Rizzo G, et al. Doppler velocimetry
1995 of the uterine artery as a screening test for gestational hyper-
6. Brosens I, Dixon HG, Robertson WB. Fetal growth retardation tension. Ultrasound Obstet Gynecol 1993;3:18–22
and the arteries of the placental bed. Br J Obstet Gynaecol 23. North RA, FerrierC, Long D, et al. Uterine artery Doppler flow
1977;84:655–63 velocity waveforms in the second trimester for the prediction
7. Meekins JW, Pijnenborg R, Hanssens M, et al. A study of of preeclampsia and fetal growth retardation. Obstet Gynecol
placental bed spiral arteries and trophoblastic invasion in 1994;83:378–86
normal and severe pre-eclamptic pregnancies. Br J Obstet 24. Todros T, Ferrazzi E, Arduini D, et al. Performance of Doppler
Gynaecol 1994;101:669–74 ultrasonography as a screening test in low risk pregnancies:
8. De Wolf F, Robertson WB, Brosens I. The ultrastructure results of a multicentric study. J Ultrasound Med 1995;
of acute atherosis in hypertensive pregnancy. Am J Obstet 14:343–8
Gynecol 1975;123:164–74 25. Harrington K, Cooper D, Lees C, et al. Doppler ultrasound
9. Sheppard BL, Bonnar J. The ultrastructure of the arterial
For personal use only.
88
J ournal of Maternal–Fetal and Neonatal Medicin e