Middle East Fertility Society Journal (2012) 17, 275282

Middle East Fertility Society

Middle East Fertility Society Journal

www.mefsjournal.org
www.sciencedirect.com

ORIGINAL ARTICLE

Prediction of fetal anemia by middle cerebral artery Doppler

a,b,* a,c
S. El. Shourbagy , M. Elsakhawy

a
Fetal Medical Unit, King George Hospital, Barkely Lane, Goodmayes, Essex IG3 8YB, UK
b
Departments of Obstetric & Gynecology, Faculty of Medicine, Tanta University, Egypt
c
Departments of Diagnostic Radiology, Faculty of Medicine, Menouya University, Egypt

Received 11 March 2012; accepted 14 September 2012

Available online 30 October 2012

KEYWORDS Abstract Objectives: To assess the value of peak systolic velocity in the middle cerebral artery
Fetal anemia; (MCA) in prediction of fetal anemia as a non invasive method in non hydropic fetuses.
Doppler of peak systolic Methods: The study included 30 pregnant women with non hydropic fetuses and with known red
velocity in MCA; cell antibodies. Full ultrasound examination was done and peak velocity of systolic cerebral blood
Cordocentesis; ow in MCA was measured. If severe anemia was suspected, fetal blood sampling by cordocentesis
Red cell antibodies was performed.
Results: Thirty fetuses were examined, 22 were anemic and eight had a hemoglobin value within
a normal range. The mean MCA peak systolic velocity for fetus with the normal hemoglobin (Hb)
was 48.98 13.94 while that for the anemic fetus was 64.79 11.97 and P = 0.004.
Sensitivity of increased peak velocity of systolic blood ow in MCA for prediction of fetal anemia
was 90.5% and specicity was 78.6%.
Conclusion: Doppler of peak velocity of systolic blood ow in MCA can be reliable in predicting
anemia so delaying invasive methods until treatment (blood transfusion) is expected to be necessary.
2012 Middle East Fertility Society. Production and hosting by Elsevier B.V. All rights reserved.

1. Introduction
Maternal alloimmunization occurs when a pregnant woman
has an immunologic response to a paternally derived red-cell
antigen that is foreign to the mother and inherited by the fetus.
The antibodies may cross the placenta, bind to antigens pres-
ent on the fetal erythrocytes, and cause hemolysis, hydrops
* Corresponding author at: Departments of Obstetric & Gynecology,
Faculty of Medicine, Tanta University, Egypt.
E-mail address: shahyshorbagy70@yahoo.com (S. El. Shourbagy).
Peer review under responsibility of Middle East Fertility Society.
Production and hosting by Elsevier
fetalis, and fetal death (1). Women with rising red cell antibody
levels, with or without a history of a previously affected preg-
nancy, are usually referred to fetal medicine units for special-
ized managements (2). Survival rates can exceed 90 percent if
anemia is diagnosed and treated with intrauterine blood trans-
fusions (3).
The most accurate test to assess the degree of fetal anemia,
and thus the need for transfusion, is the fetal hemoglobin (Hb)
measurement by fetal blood sampling (4). However, the stan-
dard test to evaluate the need for fetal transfusion is serial
amniocentesis for the determination of bilirubin levels in amni-
otic uid. Hemolysis leads to the accumulation of bilirubin in
amniotic uid, so its level correlates with the severity of hemo-
lysis. The bilirubin level is quantied by spectrophotometry
and expressed as the change in optical density at a wavelength

1110-5690 2012 Middle East Fertility Society. Production and hosting by Elsevier B.V. All rights reserved.
http://dx.doi.org/10.1016/j.mefs.2012.09.003
276
of 450 nm (DOD450); the DOD450 values are then plotted on a
Liley chart to estimate the severity of anemia (5). Both proce-
dures have a signicant rise of causing miscarriage, premature
rupture of membrane, preterm delivery and increase antibody
concentration (6).
Several non-invasive methods for the prediction of fetal
anemia have been assessed, aiming to defer the use of invasive
testing until transfusion is necessary (2,7,10). The association
of increased blood ow velocity in the fetal middle cerebral ar-
tery (MCA) with fetal anemia is the most promising of these
non-invasive tests as has subsequently been reported (9,8). Fe-
tuses with anemia have a high cardiac output and decreased
blood viscosity, resulting in high blood-ow velocities that
could be used in prediction of fetal anemia (11,12). Brennad
showed that in anemic fetus changes in cardiac output &
blood-ow velocities are reected by an increase in peak sys-
tolic velocity in the middle cerebral artery (13).
The objective of the current study is to assess the value of
measuring peak systolic velocity in the middle cerebral artery
(MCA) as a non invasive method in non hydropic fetuses
for prediction of fetal anemia due to maternal red-cell
alloimmunization.
2. Methods
2.1. Subjects studies
The study included 30 pregnant women with red cell antibodies
referred for assessment and management. Patients were mainly
referred if there was an increase in antibody levels in the mater-
nal serum, or if there was a previous obstetric history of an ane-
mic fetus or neonate. Assessment consisted of ultrasound
examination for fetal biometry, exclusion of any structural
abnormalities, liquor volume assessment, umbilical artery
Doppler blood ow studies and Doppler blood ow velocity
studies of the MCA. Hydrops was dened as uid collection
in body cavities or skin edema, and hydropic fetuses were ex-
cluded from the study. Cordocentesis was performed on 18 pa-
tients at their rst visit and for the remaining, follow- up was
arranged and those whose assessment was suggestive of wors-
ening anemia underwent cordocentesis on the second visit.
2.2. Doppler studies
Doppler examination of the MCA was performed before cor-
docentesis in all cases. An axial section of the brain, including
the thalami and the cavitas septi pellucidi, was obtained. The
circle of Willis was visualized and the middle cerebral artery
of one side was examined close to its origin in the internal car-
otid artery. We have found that the systolic velocity decreases
with distance from the point of origin of this vessel, so the angle
between the ultrasound beam and the direction of blood ow
was kept as close as possible to 0 and in all cases not to exceed
30. The highest point of the wave from peak systolic velocity
(PSV) was measured. Doppler images were recorded at a time
when there was an absence of marked fetal body and respira-
tory movements, to avoid false elevation in PSV with fetal heart
rate acceleration. At least three measurements were taken and
the highest one is accepted as the nal value. The reference test
for the diagnosis of fetal anemia was measurement of peak sys-
tolic velocity of the MCA more than 1.5 multiple of median
S. El. Shourbagy, M. Elsakhawy
(MOM) using reference produced by Mari and his colleagues
(9). Practical approach is to convert actual values into MOM
to account for changes in gestational age, internet based calcu-
lator is available at www.perinatology.com.
2.3. Cordocentesis
Fetal blood sampling was performed by cordocentesis from the
umbilical vein at the placental cord insertion with blood ready
for fetal transfusion if necessary. Fetal blood was immediately
examined for the determination of the full blood count and, if
anemia was conrmed, intravascular transfusion was
performed.
The reference test for the diagnosis of fetal anemia was
measurement of the hemoglobin level in umbilical-cord blood.
Fetal anemia is dened as Hb less than 0.65 times the median
for gestational age, with the use of the published reference
range (9).
2.4. Statistics
Statistical presentation and analysis of the present study was
conducted, using the mean by SPSS V.16.
Mean value X:  the sum of all observations divided by the
number of observations:
P
 x
X
n
Where R = sum & n = number of observations.
ROC-curve: Receiver Operating Characteristic curve
analysis
Sensitivity: Probability that the test results will be positive
when the disease is present (true positive rate, expressed as a
percentage).
Specicity: Probability that the test results will be negative
when the disease is present (true negative rate, expressed as a
percentage).
3. Results
This study included 30 pregnant women with alloimmuniza-
tion and non-hydropic fetuses. Twenty three cases had RhD
antibodies. Other 5 cases were with Rh C antibodies and last
2 cases were with anti E antibodies.
Medium gestational age was 29.27 3.54 (range 24
35 weeks). Medium maternal age was 31.1 5.67 (range 22
44 years). All patients were multi gravid. Among the 30 fetuses
at risk of anemia, 22 were anemic (Hb 8.07 0.93) and 8 had
normal Hb (Hb 10.08 0.96).
Fetal blood was taken by cordocentesis for hemoglobin
analysis in 22 cases, 18 at rst visit and four at second visit.
In the other 8 cases the MCA Doppler was within normal
range (48.98 13.94) and follow up antibodies titer remains
stable and hemoglobin was determined in fetal cord blood ob-
tained at birth and show normal Hb or mild anemia.
The mean MCA peak systolic velocity for fetuses with nor-
mal Hb was 48.98 13.94 while that for anemic fetuses was
64.79 11.97 (P 6 0.004). The MOM of MCA for fetuses
with normal Hb was 1.22 0.21 while that for anemic fetuses
was 1.62 0.08 (P 6 0.001).
There were non signicant correlations between fetal Hb
and maternal age with P = 0.19 (Fig. 1), while there were
Prediction of fetal anemia by middle cerebral artery Doppler 277

50

r.0.270
p. 0.129

Maternal age years

40

30

20
6 7 8 9 10 11 12
Fetal Hb

Figure 1 Correlation between fetal Hb and maternal age in fetuses at risk of anemia due to red-cell alloimmunization. Open rectangle
indicates fetuses with anemia & solid rectangle indicates fetuses with no or mild anemia.

50
r.0.756
p. 0.001
Maternal age years

40

30

20
30 40 50 60 70 80 90
MCA

Figure 2 Correlation between fetal peak systolic velocity of MCA and maternal age in fetuses at risk of anemia due to red-cell
alloimmunization. Open rectangle indicates fetuses with anemia & solid rectangle indicates fetuses with no or mild anemia.

1.0

.8
MOM of fetal Hb

.6

.4

.2 r. -0.358
p. 0.028

0.0
22 24 26 28 30 32 34 36
gestational age

Figure 3 Correlation between MoM of fetal Hb and gestational age in fetuses at risk of anemia due to red-cell alloimmunization. Open
rectangle indicates fetuses with anemia & solid rectangle indicates fetuses with no or mild anemia.
278 S. El. Shourbagy, M. Elsakhawy

90

80

70

MCA-PSV
60

50

40

30
22 24 26 28 30 32 34 36
gestational age

Figure 4 Correlation between MCA-PSV and gestational age in fetuses at risk of anemia due to red-cell alloimmunization. Open
rectangle indicates fetuses with anemia & solid rectangle indicates fetuses with no or mild anemia.

1.8

1.6
MOM of MCA

1.4

1.2
r.-0.574
p. 0.003

1.0
0.0 .2 .4 .6 .8 1.0
MOM of fetal Hb

Figure 5 Strong correlation between MoM of fetal Hb and MoM of MCA in fetuses at risk of anemia due to red-cell alloimmunization.
Open rectangle indicates fetuses with anemia & solid rectangle indicates fetuses with no or mild anemia.

2.0

1.8
22

1.6
MOM of MCA

1.4

1.2

1.0

.8
N= 30 22 8
Anemic fetuses Non-anemic fetuses

Figure 6 The distribution of MOM of MCA values in the anemic and non-anemic fetuses. Non anemic fetuses median 1.19, range 1.53,
maximum value 1.37, minimum value 1.03, anemic fetuses median 1.62, range1.53 maximum value 1.78, minimum value 1.52 .
Prediction of fetal anemia by middle cerebral artery Doppler 279

ROC Curve
1.00

.75

Sensitivity
.50

.25

0.00
0.00 .25 .50 .75 1.00
1 - Specificity
ROC-curve: Characteristic curve analysis for sensitivity and specificity.

Figure 7 Normal Doppler imaging of the middle cerebral artery peak systolic velocity (46.8) at 32 weeks 4 days gestation in a multi
gravida 25 years with fetal Hb10.3.

signicant correlations between MCA and maternal age with
P = 0.001 (Fig. 2).
On the other hand there was a strong negative correlation
between gestational age and MOM of fetal Hb (correlation
coefcient r = 0.358, P 6 0.028) (Fig. 3). Also, there was a
signicant correlation between gestational age and MCA (cor-
relation coefcient r = 0.525, P 6 0.014) (Fig. 4).
The relation between the multiples of the median of the
hemoglobin concentration and the multiples of the median
of peak systolic velocity was strong (correlation coefcient
r = 0. 574, P 6 0.003 for MOM of PSV) (Fig. 5).
Fig. 6 demonstrates the distribution of MOM of MCA val-
ues in the anemic and non-anemic fetuses, showing the median,
range and maximumminimum values for each group.
Sensitivity of increased PSV of MCA to predict fetal ane-
mia in cases alloimmunized with red cell antibodies was
90.5% and specicity was 78.6% (ROC Curve).
Doppler imaging of the middle cerebral artery, peak sys-
tolic velocity shows a normal level in the non anemic fetus
(where MCA was 46.8 at 32 weeks 4 days gestation in multi
gravida 25 years with fetal Hb 10.3) and was very high in cases
of fetal anemia (where MCA was 68.8 at 30 weeks 6 days
280 S. El. Shourbagy, M. Elsakhawy

Figure 8 High Doppler imaging of the middle cerebral artery peak systolic velocity (68.5) in anemic fetus with fetal Hb 7.5 at 30 weeks
6 days gestation of a multi gravida 29 years.

Figure 9 High Doppler imaging of the middle cerebral artery peak systolic velocity (74.0) in anemic fetus with fetal Hb 8.5 at 32 weeks
3 days gestation in a multi gravida 44 years.
Prediction of fetal anemia by middle cerebral artery Doppler
gestation in multi gravida 29 years with fetal Hb 7.5 & MCA
was 79.2 at 32 weeks 3 days gestation in multi gravida 44 years
with fetal Hb 8.5) (Figs. 79).
4. Discussion
Doppler velocimetry of the middle cerebral artery (MCA) has
played a major role in the fetal medicine for the last 23 years,
both in intrauterine growth-restricted (IUGR) and anemic fe-
tuses. Its utility in the diagnosis and management of cases of
fetal anemia was initially demonstrated in the cases of red cell
alloimmunization and later extended to other types of anemia
(16,15).
Data of the present ndings conrm that MCA peak sys-
tolic velocity compared with hemoglobin levels at either fetal
blood sampling, or cord sampling at delivery if antenatal inter-
vention had not been indicated is signicantly increased in the
cases of fetal anemia owing to red cell antibodies. The most
likely physiological explanation for this is that anemia is asso-
ciated with both increased cardiac output, due to the hyperdy-
namic circulation, and a reduction in blood viscosity, both
leading to increased blood ow velocity.
The association between increased MCA blood ow veloc-
ity and fetal anemia has been previously demonstrated using
peak systolic velocity (9,10,17). The use of published, generally
agreed, reference range was preferred for this study.
The peak systolic velocity in the middle cerebral artery de-
creases when the fetal hematocrit rises (20). These ndings
indicate that there is a reciprocal relation between the hemo-
globin concentration and hematocrit values and the velocity
of cerebral blood ow (18). We found that the risk of anemia
was high in fetuses with a peak systolic velocity of 1.50 times
the median or higher. Such ndings support the previous work
of Moise who found that an elevated peak MCA velocity of
>1.5 multiples of the median is useful in the timing of the
initial intrauterine transfusion (IUT) in the red cell-alloimmu-
nized pregnancies (19). Fetuses with values below 1.50 either
did not have anemia nor had only mild anemia. The fact that
this test does not predict mild anemia well is not clinically
important, because no intervention is indicated in fetuses with
mild anemia, as dened in our study, whereas those with
moderate or severe anemia should undergo cordocentesis
and may need transfusion as had been previously reported
by (1).
Although cordocentesis allows direct measurement of fetal
hemoglobin, it is associated with infection, bleeding, fetal bra-
dycardia, premature rupture of the membranes and pregnancy
loss (21). However, amniocentesis is less invasive than cordo-
centesis but the reliability of measurements of bilirubin in
amniotic uid before 27 weeks of gestation is questionable.
Yet for both procedures there are no data concerning optimal
accuracy (22,23). Opekes and his colleagues showed through
comparing Doppler ultrasonography versus amniocentesis
that measurements of the peak velocity of systolic blood ow
in MCA can safely replace invasive testing managements of
Rh-alloimmunized pregnancies (24).
Therefore, the use of measurements of peak systolic velocity
as described here would be benecial and decrease the number
of fetuses subjected to cordocentesis and amniocentesis. These
ndings coincided with the work of Alshimmiri and his
colleagues who have shown that fetal MCA-PSV weakly
281
correlates with the degree of fetal anemia in groups with or
without previous intrauterine transfusion; however, velocity
threshold levels were sensitive enough to predict most fetuses
with moderate to severe anemia (25).
Our ndings coincided with the work of previous research-
ers who demonstrated that MCA-PSV measurement is essen-
tial in the diagnosis, evaluation, and management of the
cases of fetal anemia (26,27). The use of this modality lessens
the need for invasive procedures. Also, Kenneth and Moise
showed that the peak systolic velocity of the MCA was effec-
tive in the detection of fetal anemia in a variety of pathologic
states (28). They showed that addition of this noninvasive
method to detect the anemic fetus has enabled the maternal-fe-
tal specialist to intervene earlier in the course of such diseases
as hemolytic disease of the fetus/newborn infant and fetal par-
vovirus infection. Although the middle cerebral artery peak
systolic velocity Doppler has limited diagnostic accuracy, it re-
mains the gold standard for noninvasive screening of fetal ane-
mia (29).
Our results were compared with hemoglobin levels at either
fetal blood sampling in 22 cases, or cord sampling at delivery
in 8 cases where antenatal intervention had not been indicated.
MCA-PSV had a sensitivity of 90.5% and a specicity of
78.6% for the detection of severe anemia. Brennand showed
by comparing hemoglobin levels in 165 fetuses at either fetal
blood sampling, or cord sampling at delivery that the sensitiv-
ity and accuracy of the middle cerebral artery Doppler (sensi-
tivity of 88%, specicity of 82%) were substantially greater
than amniocentesis (sensitivity of 76%, specicity of 77%)
for the detection of anemia (13).
Hydropic fetuses were excluded from this study. This is be-
cause we wished to assess how effective the measurement of
MCA Doppler velocity is at predicting fetal anemia in border-
line cases, in which the decision to sample the fetal blood is not
clear.
In conclusion, measurements of the peak velocity of blood
ow in the middle cerebral artery in fetuses at risk for anemia
due to maternal red cell alloimmunization provide an accurate
and noninvasive means of determining the degree of anemia.
Clearly, the widespread use in future of MCA Doppler assess-
ment to detect fetal anemia in other fetal diseases is on the near
horizon. However, no non-invasive test is 100% accurate in all
cases, and so the assessment of patients with red cell antibodies
should be comprehensive and include all possible relevant
information, including obstetric history and rate of change
of antibody levels.
References
(1) Bowman JM. Haemolytic disease (Erythroblastsis fetalis). In:
Creasy RK, Rensnik R, editors. Maternal-fetal medicine 4th.
Philadelphia: W.B. Saunders; 1999. p. 73667.
(2) Oepkes D, Brand R, Vandenbussche FP, Meerman RH, Kanhai
HUH. The use of ultrasonography and Doppler in the prediction
of fetal haemolytic anaemia: a multivariate analysis. Br J Obstet
Gynaecol 1994;101:6804.
(3) van Kamp IL, Klum FJCM, Bakkum RSI. The severity of
immune fetal hydrops is predictive for fetal outcome after
intrauterine treatment. Am J Obstet Gynaecol 2001;185:66873.
(4) Soothill PW. Fetal blood sampling before labor. In: James DK,
Steer PJ, Weiner CP, Gonik B, editors. High risk pregnancy,
management options. London: WB Saunders; 1999. p. 22533.
282
(5) Nicolaides KH, Rodeck CH, Mibashan RS, Kemp JR. Have
Liley charts outlived their usefulness? Am J Obstet Gynaecol
1986;155:6804.
(6) Weiner CP, Okamura K. Diagnostic fetal blood sampling-
technique related losses. Fetal Diagn Ther 1996;11(3):16975.
(7) Hadley AG, Wilkes A, Goodrick J, Penman DW, Lucas G. The
ability of the chemiluminescence test to predict clinical outcome
and the necessity for amniocenteses in pregnancies at risk of
haemolylic disease of the newborn. Br J Obstet Gynecol 1998;105:
2314.
(8) Delle Chiaie L, Buck G, Grab D, Terinde R. Prediction of fetal
anemia with Doppler measurement of the middle cerebral artery
peak systolic velocity in pregnancies complicated by maternal
blood group alloimmunization or parvovirus B19 infection.
Ultrasound Obstet Gynecol 2001;18(3):2326.
(9) Mari G, Deter RL, Carpenter RL. Noninvasive diagnosis by
Doppler ultrasonography of fetal anemia due to maternal red-cell
alloimmunization. N Eng J Med 2000;342:914.
(10) Teixeira IMA, Duncan K, Letsky E, Fisk NM. Middle cerebral
artery peak systolic velocity in the prediction of fetal anemia.
Ultrasound Obstet Gynaecol 2000;I5:3058.
(11) Duker D, Oepkes D, Seaward G, Windrim R, Ryan G.
Noninvasive tests to predict fetal anemia: a study comparing
Doppler and ultrasound parameter. Am J Obstet Gynaecol 2003;
188:13104.
(12) Segata M, Mari G. Fetal anemia: new technologies. Curr Opin
Obstet Gynaecol 2004;16:1538.
(13) Brennand J. Middle cerebral artery Doppler. AJUM
2009;12(4):359.
(15) Schenone MH, Mari G. The MCA Doppler and its role in the
evaluation of fetal anemia and fetal growth restriction. Clin
Perinatol 2011;38(1):83102.
(16) Hanif F, Drennan K, Mari G. Variables that affect the middle
cerebral artery peak systolic velocity in fetuses with anemia and
intrauterine growth restriction. Am J Perinatol 2007;24(8):5015.
(17) Ahmed B, Ghaffari Z, Ismail RS, Saleh N. Non-invasive
diagnosis of fetal anemia due to maternal red-cell alloimmuniza-
tion. Saudi Med J 2005;26(2):2569.
(18) Mandic V, Mikovic Z, Filimonovic D, Dukic M, Vasiljevic M,
Bogavac M. Noninvasive diagnosis of fetal anemia in Rh-alloim-
munized pregnancies by color Doppler. Med Pregl 2005;58(56):
2758.
S. El. Shourbagy, M. Elsakhawy
(19) Moise Jr KJ. The usefulness of middle cerebral artery Doppler
assessment in the treatment of the fetus at risk for anemia. Am J
Obstet Gynecol 2008;198(2):161.e14.
(20) Mari G, Rahman F, Olofsson P, Ozcan T, Copel JA. Increase
fetal hematocrit decreases middle cerebral artery peak systolic
velocity in pregnancies complicated by rhesus alloimmunization. J
Matern Fetal Med 1997;6:2068.
(21) Ghidini A, Sepulveda W, Lockwood CJ, Romero R. Complica-
tion of fetal blood sampling. Am J Obstet Gynaecol 1993;168:
133944.
(22) Rahman F, Detti L, Ozcan T, Khan R, Manohar S, Mari G. Can
a single measurement of amniotic uid delta optical density be
safely used in the clinical management of rhesus-alloimmunized
pregnancies before 27 weeks gestation? Acta Obstet Gynecol Scan
1998;77:8047.
(23) Bullock R, Martin WL, Coomarasamy A, Kilby MD. Prediction
of fetal anemia in pregnancies with red -cell isoimmunization
comparison of middle cerebral artery peak systolic velocity and
amniotic uid DOD450. Ultrasound Obstet Gynaecol 2005;25:
3314.
(24) Oepkes D, Seaward G, Vandenbussche FP, Windrim R, Beyene J,
Kanhai HUH. Doppler ultrasonography versus amniocentesis to
predict of fetal anaemia. N Eng J Med 2006;355:15664.
(25) Alshimmiri MM, Hamoud MS, Al-Saleh EA, Mujaibel KY, Al-
Harmi JA, Thalib L. Prediction of fetal anemia by middle cerebral
artery peak systolic velocity in pregnancies complicated by rhesus
isoimmunization. J Perinatol 2003;23:53640.
(26) Imbar T, Lev-Sagie A, Cohen S, Yanai N, Yagel S. Diagnosis,
surveillance, and treatment of the anemic fetus using middle
cerebral artery peak systolic velocity measurement. Prenat Diagn
2006;26(1):4551.
(27) Lubusky M, Prochazka M, Santavy J, M ckova I, Kantor L.
Actual management of pregnancies at risk for fetal anemia. Ceska
Gynekol 2006;71(4):27280 [Article in Czech].
(28) Kenneth J, Moise Jr. The usefulness of middle cerebral artery
Doppler assessment in the treatment of the fetus at risk for
anemia. Am J Obstet Gynecol 2008;198:161.e14.
(29) Pretlove SJ, Fox CE, Khan KS, Kilby MD. Noninvasive methods
of detecting fetal anaemia: a systematic review and meta-analysis.
BJOG 2009;116(12):155867.