Ultrasound Obstet Gynecol 2009; 33: 209 212

Published online in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/uog.6301

Cervical length and risk of antepartum bleeding in women

with complete placenta previa
T. GHI*, E. CONTRO*, T. MARTINA*, M. PIVA*, R. MORANDI*, L. F. ORSINI*,
M. C. MERIGGIOLA*, G. PILU*, A. M. MORSELLI-LABATE, D. DE ALOYSIO*, N. RIZZO*
and G. PELUSI*
Departments of *Obstetrics and Gynecology and Internal Medicine and Gastroenterology, University Hospital of Bologna, Bologna, Italy

K E Y W O R D S: antepartum bleeding; cervical length; placenta previa; preterm delivery; ultrasound

AB S T RA C T
Objective To evaluate if cervical length predicts prepartum bleeding and emergency Cesarean section in cases of
placenta previa.
Methods Between September 2005 and September 2007,
cervical length was measured by transvaginal ultrasound
in women with complete placenta previa persisting into
the third trimester of pregnancy. A complete follow-up of
pregnancy was obtained in all cases.
Results Overall, 59 women were included in the study
group. The mean SD gestational age at ultrasound was
30.7 2.7 weeks and the cervical length was 36.9
8.8 mm. Cesarean delivery was performed in all cases,
at a mean gestational age of 34.7 2.3 weeks. Twentynine (49.1%) of the women presented prepartum bleeding
and
12 (20.3%) required an emergency Cesarean section prior
to 34 completed weeks due to massive hemorrhage. Cervical length did not differ significantly between cases
with and those without prepartum bleeding (35.3 9.3
mm vs.
38.4 8.2 mm; P = 0.18), but was significantly
shorter among patients who underwent emergency
Cesarean sec- tion < 34 weeks
due to massive
hemorrhage compared with patients who underwent
elective Cesarean section (29.4 5.7 mm vs. 38.8 8.5
mm; P = 0.0006).
Conclusions Transvaginal sonographic cervical length
predicts the risk of emergency Cesarean section
< 34 weeks in women with complete placenta previa.
Copyright
2009 ISUOG. Published by John Wiley
& Sons, Ltd.

IN TR O D UC TI O N

rising rates of Cesarean section. It is estimated to occur

in between 0.2 and 0.3% of third-trimester pregnancies1 .
A conclusive diagnosis of complete placenta previa is
ascertained sonographically in the third trimester, when
upward migration from the internal cervical os becomes
unlikely2 .
Women with placenta previa are at increased risk of
prepartum maternal bleeding and emergency preterm
Cesarean section leading to perinatal complications.
Furthermore, in these cases the higher chance of
postpartum complications, including uterine atony and
placenta accreta, is well established. However, the risk of
maternal hemorrhage or prematurity is unpredictable.
In theory, such complications could be expected more
frequently in those patients whose risk of preterm labor
is also increased. An inverse relationship between cervical
length measured by transvaginal ultrasound and the
risk of spontaneous preterm labor has been clearly
demonstrated3,4 . The aim of our study, therefore, was
to determine if transvaginal ultrasound assessment of
cervical length predicts the risk of prepartum bleeding
or preterm emergency Cesarean section due to massive
hemorrhage in women with complete placenta previa
diagnosed in the third trimester of pregnancy.

PATI E N TS A ND M ET HOD S
Between September 2005 and September 2007 all
asymptomatic women
diagnosed consecutively with
complete placenta previa at the ultrasound laboratory
of our University Hospital were selected for the purpose
of this study. Patients had been referred to our center due
to sonographic suspicion of placenta previa. Diagnosis
of complete placenta previa was made by transvaginal
ultrasound in the third trimester of pregnancy when the
lower placental edge appeared to overlay completely the

Placenta previa has been diagnosed increasingly in recent

decades, due to the widespread use of ultrasound and the

Correspondence to: Dr T. Ghi, I Clinica Ostetrica-Ginecologica, Policlinico S. Orsola-Malpighi, Via Massarenti 13, 40100 Bologna, Italy
(e-mail: tullio.ghi@aosp.bo.it)
Accepted: 22 September 2008
Copyright

2009 ISUOG. Published by John Wiley & Sons, Ltd.

ORIGINAL PAPE
R

Ghi et al.

210
internal os of the uterine cervix. We excluded patients
with any of the following conditions: gestational age
< 27.0 and 36.0 completed weeks; multiple pregnancy;
threatened preterm labor
or premature rupture of
membranes; history of bleeding in the current pregnancy;
polyhydramnios; history of cervical cone biopsy; presence
of cerclage; maternal use of vaginal
progesterone;
sonographic suspicion of fetal anomaly or fetal growth
restriction; history of maternal disease or hypertensive
disorder complicating the pregnancy.
In all women included in the study group, cervical
length was measured transvaginally at the time of
diagnosis. Ultrasound examination was performed using
a machine equipped with a multifrequency transvaginal
probe (Tecnos and MyLab 50 Xview, Esaote, Genoa,
Italy). Cervical evaluation was performed according to a
standardized technique: women were asked to void their
bladder before the examination. A true sagittal plane
was obtained in order to visualize the full length of the
cervical canal and cervical length was measured three
times by placing the callipers on the internal and external
os. The shortest measurement was then recorded.
Following sonographic diagnosis of complete placenta
previa, induction of lung maturity was carried out and
elective Cesarean section was scheduled between 36
and 37 completed weeks of gestation. In all cases, low
placental insertion was confirmed sonographically just
prior to Cesarean delivery. If vaginal bleeding occurred
prior to the scheduled Cesarean section, patients were
admitted to the hospital and timing of delivery was
decided according to the clinical condition of the patient.
If the patient was clinically stable, delivery was carried
out electively beyond 34 weeks of gestation, following
documentation of fetal lung maturity by amniocentesis.
In cases of massive maternal hemorrhage, emergency
Cesarean section was performed before 34 weeks of
gestation. Surgeons and obstetricians who assisted at the
delivery were blinded to cervical length. Cases in which the
indication for emergency Cesarean section was something
other than massive hemorrhage or in which placenta
accreta was diagnosed intraoperatively were excluded
from the study.
A complete follow-up of pregnancy was obtained in
all cases, with the following variables being recorded:
gestational age at delivery, neonatal weight, occurrence of
prepartum bleeding, type of Cesarean delivery (elective vs.
emergency Cesarean section due to massive hemorrhage).
Mean cervical lengths in the groups with and without
prepartum bleeding
were compared retrospectively.
Furthermore, we compared cervical length between
patients who underwent elective Cesarean section and
those who underwent emergency Cesarean section due to
massive hemorrhage.

Statistical analysis
Data are reported as means and SDs. Continuous
variables (cervical lengths) were compared using Students
t-test. A two-tailed value of P < 0.05 was considered

Copyright

2009 ISUOG. Published by John Wiley & Sons, Ltd.

statistically
significant.
Receiver

operating
characteristics (ROC)
curves were constructed to
determine the accuracy of cervical
length
for the
identification of women at high risk for prepartum
vaginal bleeding and emergency Cesarean section < 34
weeks due to massive hemorrhage. The areas under the
ROC curves together with their standard error were
evaluated. The best cut-off was evaluated according to
maximum likelihood ratio. The SPSS Version 13.0 (SPSS
Inc., Chicago, IL, USA) statistical package was used to
analyze data.

RESULTS
A total of 60 asymptomatic women with complete
placenta previa were enrolled prospectively and included
in the study group. A further five more cases were
considered ineligible due to previous bleeding in the
current pregnancy. A case of placenta accreta was
excluded retrospectively, leading to a final population
of 59 women. A summary of patient data and pregnancy
outcome is provided in Table 1. Twenty-nine women
(49.1%) were nulliparous, parity was 1 in 21 women
and it was more than one in nine women. The
mean SD gestational age at transvaginal ultrasound
was 30.7 2.7 weeks and the cervical length was
36.9 8.8 mm. Cervical length did not vary significantly
between nulliparous and parous women (37.9 8.8 mm
vs. 35.9 8.9 mm; P = 0.375). Cesarean section was
performed in all cases, at a mean gestational age of 34.7
2.3 weeks. The mean birth weight was 2440.5 497.6 g.
Twenty-nine (49.1%) of the women presented antepartum vaginal bleeding, 12 (20.3%) of whom required an
emergency Cesarean section prior to 34 completed weeks
due to massive hemorrhage (Table 2). There were no cases
of emergency Cesarean section due to massive hemorrhage
after 34 completed weeks of gestation. All women without prepartum bleeding reached the scheduled date for
Table 1 Patient characteristics and pregnancy outcome in 59
women with complete placenta previa in the third trimester of
pregnancy
Characteristic/outcome
Maternal age (years)
Parity
0
1
2
>2
Previous Cesarean section
0
1
>1
Gestational age at diagnosis (weeks)
Cervical length (mm)
All women
Nulliparous women
Parous women
Gestational age at delivery (weeks)
Birth weight (g)

n (%) or mean SD
34.9 4.5
29 (49.1)
21 (35.5)
6 (10.1)
3 (5.1)
11 (18.6)
12 (20.3)
7 (11.8)
30.7 2.7
36.9
37.9
35.9
34.7
2440.5

8.8
8.8
8.9
2.3
497.6

Ultrasound Obstet Gynecol 2009; 33: 209 212.

Cervical length and bleeding in placenta previa

211

Table 2 Outcome of pregnancy according to cervical length in 59

women with complete placenta previa in the third trimester of
pregnancy

Vaginal bleeding
With
Without
Cesarean section
Emergency < 34 weeks
Elective 34 weeks

n (%)

Cervical length
(mean SD)

0.8
P
0.18

29 (49.1)
30 (50.8)

35.3 9.3
38.4 8.2

12 (20.3)
47 (79.6)

29.4 5.7
38.8 8.5

0.0006

Sensitivity

Outcome

1.0

0.4

elective Cesarean section beyond 36 weeks of gestation,

while in those with minor bleeding prior to the scheduled
delivery, Cesarean section was performed electively after
34 weeks following documentation of fetal lung maturity.
In no case was the indication for emergency Cesarean

DI SC U S SI O N
Our study has confirmed that women with placenta
previa are at higher risk of complications, including
prematurity and severe hemorrhage. In this series, the
chance of prepartum bleeding among women diagnosed
sonographically with placenta previa in the third trimester
was approximately 50%, while the risk of emergency
Cesarean section prior to 34 completed weeks due to

0.2

0.2

0.4

0.6

0.8

1.0

1 Specificity
Figure 1 Receiver operating characteristics curve for cervical
length in the prediction of vaginal bleeding in women with
complete placenta previa in the third trimester of pregnancy.
1.0

0.8

Sensitivity

section something other than massive hemorrhage.

The mean gestational age at delivery (32.5
1.4 weeks vs. 35.3 2.1 weeks; P = 0.00005) and mean
birth weight (2136.6 356.6 g vs. 2518.1 501.5
g; P = 0.016) were significantly lower in women who
under- went emergency Cesarean section compared with
those who underwent elective Cesarean section. The
occur- rence of prepartum bleeding and emergency
section were comparable among nulliparous and parous
women (OR, 1.46 (95% CI, 0.42 5.01) vs. 1.41
(95% CI,
0.51 3.87)). Cervical length did not differ significantly
between cases with and those without prepartum bleeding
(35.3 9.3 mm vs. 38.4 8.2 mm; P = 0.18), but
was significantly shorter in patients who underwent
emergency Cesarean section < 34 weeks due to massive
hemorrhage compared with those who underwent
elective Cesarean section 34 weeks (29.4 5.7 mm
vs. 38.8 8.5 mm; P = 0.0006) (Table 2). ROC curves
for cervical length in the prediction of vaginal bleeding
and for cervical length in the prediction of emergency
Cesarean section < 34 weeks are given in Figures 1 and 2,
respectively. The best cut-off point for the identification of
women at high risk for emer- gency Cesarean section <
34 weeks as determined from
the ROC curve (Figure 2) was cervical length 31 mm
(with 83.3% sensitivity, 76.6% specificity, 47.6% positive predictive value (PPV) and 94.7% negative predictive
value (NPV) (P < 0.001). In particular, women whose
cervical length was 31 mm at sonographic diagnosis of
placenta previa had a 16 times higher risk of preterm
Cesarean delivery due to massive hemorrhage (OR, 16.36
(95% CI, 3.39 75.92)).

0.6

0.6

0.4

0.2

0
0

0.2

0.4

0.6

0.8

1.0

1 Specificity
Figure 2 Receiver operating characteristics curve for cervical
length in the prediction of emergency Cesarean section < 34 weeks
in women with complete placenta previa in the third trimester of
pregnancy.

massive hemorrhage was almost one in five. This latter

risk seemed to be higher among women whose cervical
length was shorter. This observation may be of clinical
value if our data are confirmed in larger series.
Placenta previa is among the most frequently diagnosed
obstetric conditions, with an increasing prevalence due to
the widespread use of ultrasound and the rising rates of
Cesarean section. However, women who are more prone
to developing severe bleeding requiring an emergency
premature delivery are not likely to be recognised in
the preclinical stage. Consequently, all asymptomatic

Ghi et al.

212
women with complete placenta previa detected in the
third trimester are usually scheduled for elective Cesarean
section around 36 weeks of gestation unless vaginal
bleeding occurs earlier. However, obstetric management
is not tailored specifically to the patient because individual
risk of hemorrhage is difficult to predict. Our study
seems to suggest that the shorter the cervix at the time
of sonographic diagnosis of complete placenta previa
in the third trimester, the higher the risk of severe
prematurity and emergency Cesarean section due to
massive maternal hemorrhage. In our series, if cervical
length was 31 mm, the risk of preterm Cesarean section
was almost one in two, whereas, due to the excellent
negative predictive value of the ultrasound examination,
whenever the cervical length was above this threshold, the
need for preterm Cesarean section was rare.
We controlled for possible confounding factors which
might have contributed significantly and independently to
the outcomes, particularly the risks of vaginal bleeding
and preterm delivery. Parity did not seem to interfere
with the outcome, occurrence of prepartum bleeding
or emergency Cesarean section not varying significantly
between nulliparous and parous women. Furthermore, we
excluded retrospectively from the study a case of placenta
accreta, as hemorrhage due to detachment would have
been less likely in this case.
The main limitation of our study is the small number of
women enrolled, but, if our observations are confirmed in
larger series, transvaginal measurement of cervical length
could become part of the routine third-trimester scan in
women with complete placenta previa in order to predict
the risk of severe complications.
The increased risk of preterm hemorrhage in women
with complete placenta previa and short cervix could be
explained by the fact that with a short cervix the chance
of spontaneous preterm labor is increased. Sonographic
detection of a short cervix has been demonstrated
consistently to predict earlier occurrence of labor5,6 .
A shorter cervix among women with placenta previa
may therefore also herald premature onset of labor and
possible detachment of the placenta from its low insertion.
Classically, hemorrhage from placenta previa during labor
has been related to the fact that, due to its inelastic
structure, the placenta may not adapt to progressive
cervical effacement and undergoes detachment7 . Our
study suggests that in women with placenta previa,
although the risk of minor prepartum bleeding does not
seem to vary according to cervical length, shortening of
the cervix even to a minor extent may predict an earlier
placental detachment with massive hemorrhage.
In women with placenta previa, the use of transvaginal
ultrasound has been aimed more frequently at predicting
persistence8 or adhesion9 of placenta or at detecting

possible coexistence with vasa previa7 , whereas its role in

predicting the risk of maternal bleeding has been mostly
unexplored. An increased thickness of the lower placental
edge has been reported by some to increase the risk of
prepartum hemorrhage10 . Others have shown a higher
chance of massive bleeding among cases with an echo-free
space and lacunae within the lower placental edge11 . Our
findings regarding a possible association between cervical
length and the risk of preterm hemorrhage in patients
with complete placenta previa have not been reported
previously and may improve our ability to predict clinical
course and to refine obstetric management in these cases.
If our data are confirmed on larger numbers, an earlier
hospital admission, or at least closer clinical monitoring,
in women with placenta previa and short cervix may
become a option.

R EF ERENC E S
1. Tuzovic L. Complete versus incomplete placenta praevia and
obstetric outcome. Int J Gynaecol Obstet 2006; 93: 110 117.
2. Oppenheimer L, Holmes P, Simpson N, Dabrowski A. Diagnosis of low-lying placenta: can migration in the third trimester predict outcome? Ultrasound Obstet Gynecol 2001; 18: 100 102.
3. Colombo DF, Iams JD. Cervical length and preterm labor.
Clin Obstet Gynecol 2000; 43: 735 745.
4. Iams JD, Goldenberg RL, Meis PJ, Mercer BM, Moawad A,
Das A, Thom E, McNellis D, Copper RL, Johnson F, Roberts JM.
The length of the cervix and the risk of spontaneous premature
delivery. National Institute of Child Health and Human Development Maternal Fetal Medicine Unit Network. N Engl J Med
1996; 334: 567 572.
5. Rozenberg P, Gillet A, Ville Y. Transvaginal sonographic
examination of the cervix in asymptomatic pregnant women:
review of the literature. Ultrasound Obstet Gynecol 2002; 19:
302 311.
6. Palacio M, Sanin-Blair J, Sa nchez M, Crispi F, Go mez
O, Carreras E, Coll O, Cararach V, Grataco s E. The use
of a variable cut-off value of cervical length in women
admitted for preterm labor
before and after 32 weeks.
Ultrasound Obstet Gynecol 2007; 29: 421 426.
7. Oyelese Y, Smulian JC. Placenta praevia, placenta accreta, and
vasa praevia. Obstet Gynecol 2006; 107: 927 941.
8. Oppenheimer L. Society of Obstetricians and Gynaecologists
of Canada. Diagnosis and management of placenta praevia.
J Obstet Gynaecol Can 2007; 29: 261 273.
9. Sumigama S, Itakura A, Ota T, Okada M, Kotani T, Hayakawa H,
Yoshida K, Ishikawa K, Hayashi K, Kurauchi O, Yamada S,
Nakamura H, Matsusawa K, Sakakibara K, Ito M, Kawai M,
Kikkawa F. Placenta praevia increta/percreta in Japan: a retrospective study of ultrasound findings, management and clinical
course. J Obstet Gynaecol Res 2007; 33: 606 611.
10. Ghourab S. Third-trimester transvaginal ultrasonography in
placenta praevia: does the shape of the lower placental edge
predict clinical outcome? Ultrasound Obstet Gynecol 2001; 18:
103 108.
11. Saitoh M, Ishihara K, Sekiya T, Araki T. Anticipation of uterine
bleeding in placenta praevia based on vaginal sonographic
evaluation. Gynecol Obstet Invest 2002; 54: 37 42.