European Journal of Obstetrics & Gynecology and

Reproductive Biology 119 (2005) 189193

www.elsevier.com/locate/ejogrb

Comparison of two preparations of dinoprostone for pre-induction of

labour in nulliparous women with very unfavourable
cervical condition: a randomised clinical trial
Fabio Facchinetti*, Paolo Venturini, Gianluca Verocchi, Annibale Volpe
Unit of Gynaecology and Obstetrics, Mother-Infant Department, University of Modena and Reggio Emilia,
Via del Pozzo 71, 41100 Modena, Italy
Received 11 November 2003; accepted 13 June 2004

Abstract

The aim of this study was to compare the clinical effects of preinduction cervical ripening by using two ways of dinoprostone
administration. In a prospective, open-label trial, 144 consecutive nulliparous women with a Bishop score <4 who required induction of
labour at term were randomised to receive dinoprostone via either a vaginal insert (10 mg over 12 h) or a cervical gel (0.5 mg, twice in 12 h). If
labour did not start by 24 h after this preinduction, patients received 2 mg vaginal dinoprostone gel followed 6 h later by oxytocin infusion.
The main outcome measure was the rate of caesarean sections (CS). Secondary measures were: changes in Bishop score at 6 h and 12 h,
delivery within 12 h and 24 h, need for oxytocin for induction, failure of induction (delivery after >48 h), need for pharmacological
interventions to manage tachysystole/hyperstimulation, length of stay in hospital. The CS rate was lower in the vaginal insert group (22.9%)
than in the cervical gel group (34.3%), though the difference did not reach statistical significant difference (P = 0.13). The indications for CS
overlapped between the groups. The rate of vaginal delivery within 12 h and 24 h was similar, as was the rate of failure of induction. More
women in the gel group (41.4% versus 24.3%) required the use of oxytocin (OR = 2.21; 95% CI = 1.074.55). Tachysystole or
hyperstimulation in the vaginal insert group (7) was twice that with cervical gel (4). Four women in the vaginal insert group and three
in the cervical gel group reported infectious complications. A long stay in hospital (>4 days) was less frequent with vaginal inserts (21.4
versus 38.6%; OR = 0.43, 95% CI = 0.190.97).
The more challenging preinductions of labour at term are associated with increased obstetric interventions such as a high CS rate and a
more frequent requirement for oxytocin inductions. In terms of success and failure, vaginal inserts releasing dinoprostone do not differ from
dinoprostone given by the traditional cervical route. However, the use of vaginal inserts reduces the need for obstetric interventions and allows
shorter periods in hospital.
# 2004 Elsevier Ireland Ltd. All rights reserved.

Keywords: Cervical maturation; Caesarean section; Fetal heart rate; Dinoprostone; Labour induction

1. Introduction elevated caesarean section (CS) rate [2]. A slow-release

The advantages of preinduction cervical maturation by
topical application of prostaglandins have been known for
several years [1]. However, selected groups of patients, such
as nulliparous women and those with a very unfavourable
condition of the cervix, are still a challenge to clinicians, both
these groups being associated with more failures and an
* Corresponding author.
E-mail address: facchinetti.fabio@unimore.it (F. Facchinetti).
dinoprostone vaginal insert is now available worldwide. Two
meta-analyses of the efficacy of this formulation reached
contrasting conclusions. Sanchez-Ramos et al. [3] concluded
that the vaginal insert was less effective than other
prostaglandins, whereas Hughes et al. [4] found no
clinically significant differences between the effects of this
preparation and of other standard prostaglandin preparations.
However, the clinical trials so far published involved
nonhomogeneous populations with both nulliparous and
multiparous women in most cases. In addition, the baseline

0301-2115/$ see front matter # 2004 Elsevier Ireland Ltd. All rights reserved.
doi:10.1016/j.ejogrb.2004.06.039
190 F. Facchinetti et al. / European Journal of Obstetrics & Gynecology and Reproductive Biology 119 (2005) 189193
cervical conditions were often so diverse that the Bishop
score ranged from 0 to 8 in the same patient sample. This
heterogeneity of inclusion/exclusion criteria in the available
trials seems to be the main reason why no conclusions have
been reached that can be applied to specific clinical situa-
tions.
For these reasons, we decided to compare the efficacy and
safety of the slow-release vaginal dinoprostone insert with
dinoprostone administered intracervically in a challenging
clinical population of a kind frequently encountered; we
selected that of nulliparous women with very unfavourable
cervical conditions requiring labour induction at term.
2. Materials and methods
2.1. Subjects
This prospective, randomised, controlled study was carried
out between May 2001 and June 2003 in consecutive patients
requiring labour induction. Inclusion criteria were: singleton
pregnancy, nulliparity, Bishop score < 4, gestational age >
37.0 according to a first trimester ultrasound evaluation.
Exclusion criteria were: previous uterine surgery, known
hypersensitivity to prostaglandins, fetal malpresentation,
suspected cephalopelvic disproportion, placenta praevia,
rupture of membranes, any other condition contraindicating
vaginal delivery.
One hundred and forty-four consecutive women requiring
induction of labour and fulfilling the above criteria were
randomised. The study was approved by the IRB. Written
informed consent was signed before treatment, as in every
case of labour induction at our institution.
The patients were allocated to treatment with either the
10 mg vaginal dinoprostone insert (Propess, Ferring,
Germany) or 0.5 mg intracervical dinoprostone (Prepidil
gel, Upjohn, Kalamazoo, MC, USA) in the morning in blocks
of eight, with the aid of a computer-generated random list.
The resident in charge of labour induction was informed of
the next treatment allocation by the senior midwife (who kept
the list). Treatment groups were designated the vaginal
insert and cervical gel groups. Data were collected and
analysed by one of us (P.V.), and the random assignments
were not revealed until after the analysis was complete. Then
results were presented to the other co-authors for discussion.
2.2. Protocol
The treatment protocol was as follows:
1. In the outpatient clinic, the women underwent a gynae-
cological examination and an ultrasound scan, and were
then scheduled for labour induction on the next day.
2. Patients were admitted in the morning, signed their
informed consent forms, underwent Bishop scoring and
were randomised.
3. Fetal heart rate (FHR) and uterine activity were
monitored for 30 min.
4. Either a vaginal insert or intracervical gel was then
administered.
5. FHR and uterine activity were monitored for at least
30 min every 2 h.
6. After 6 h a gynaecological examination was performed to
assess the Bishop score: in the cervical gel group women
with a Bishop score < 4 received another dose of
intracervical gel, while those with Bishop scores between
4 and 6 received 2 mg dinoprostone gel by the vaginal
route (Prepidil gel).
7. Twelve hours after the start of treatment the Bishop score
was re-evaluated and the vaginal insert was removed.
8. On the next morning, 24 h after the start of treatment
2 mg dinoprostone was administered to women in both
groups who were not yet experiencing regular uterine
contractions.
9. Six hours later every woman who was not yet in labour
had her labour induced with an oxytocin infusion given
according to a standard protocol. Amniotomy was
performed only in women in labour with at least 5 cm
of cervical dilatation. Oxytocin augmentation was done
in the second stage of labour if necessary.
2.3. Statistical analysis
The main outcome measure was the rate of caesarean
sections (CS). When intracervical gel was used for labour
induction in nulliparous women with immature cervix we
observed a CS rate of 38% (unpublished data). Data in the
literature data indicate a lower rate of CS with the vaginal
insert than with other prostaglandin preparations, and on the
basis of this we had set up the hypothesis that this vaginal
insert would allow a reduction of 50% in the CS rate in our
population. Therefore, we estimated that 138 cases were
enough to allow evaluation of the difference between
treatments with 80% power.
Secondary measures were: changes in Bishop score at 6 h
and 12 h, delivery within 12 h and 24 h, need for oxytocin to
attain induction, failure of induction (delivery >48 h after
administration), requirement of pharmacological interven-
tion to manage tachysystole/hyperstimulation, length of
admission stay.
3. Results
One hundred and forty subjects completed the study. In
four other cases (two in each group) the protocol was
violated because the mothers developed intolerance to their
contractions in the absence of tachysystole or hyperstimula-
tion. These patients refused to continue induction and then
required and underwent CS a few hours after the start of
preinduction cervical ripening. According to the intention-
to-treat analysis, demographic and clinical data in the 144
 F. Facchinetti et al. / European Journal of Obstetrics & Gynecology and Reproductive Biology 119 (2005) 189193 191

Table 1
Demographic and clinical data in the 144 patients who started on the study
Propess (N = 72) Gel (N = 72) Significance (two-tailed)
Age 29.1  4.9 27.9  5.1 N.S.
Gestational age 40.7  1.4 40.9  1.1 N.S.
Indication for labour induction
Prolonged pregnancya 33 (45.8%) 32 (44.4%) N.S.
AFI <5 24 (33.3%) 25 (34.7%)
Hypertension 13 (18.1%) 14 (19.5%)
Fetal causes 2 (2.8%) 1 (1.4%)
Preinduction Bishop score
01 40 (55.6%) 36 (50.0%) N.S.
23 32 (44.4%) 36 (50.0%)
AFI, amniotic fluid index.
a
41 weeks + 3 days or more.

Table 2
Indications for CS in the two groups
Vaginal insert (N = 70) Cervical gel (N = 70) Significance (two-tailed)
Vaginal delivery 55 (78.6%) 46 (65.7%) N.S.
CS because of
Fetal distress 4 (5.7%) 8 (11.4%)
Dystocia 6 (8.6%) 7 (10.0%)
Failed induction 5 (7.1%) 9 (12.9%)

patients who started on the study are reported in Table 1. No
significant changes were found between the two groups.
The CS rate was lower in the vaginal insert group (22.9%)
than in the cervical gel group (34.3%), though the difference
did not reach statistical significance (Chi-square: 2.24, P =
0.13). The indications for CS overlapped between the two
groups, as detailed in Table 2.
Secondary outcome measures are reported in Table 3.
The rate of vaginal delivery within 12 h and 24 h was similar
in both groups, as was the rate of failure of induction. The
interval from induction to vaginal delivery was similar in the
vaginal insert (1374  609 min) and the cervical gel (1343
 595 min) groups. More women in the cervical gel group
(41.4%, as against 24.3% in the vaginal insert group; Chi-
square3.92, P = 0.48) needed oxytocin for labour induction,
giving an OR of 2.21 (95% CI = 1.074.55).
Almost twice as many subjects in the vaginal insert group
(7) as in the cervical gel group (4) needed drugs for manage-
ment of tachysystole/hyperstimulation, but this difference
also failed to reach statistical significance. Four women in the

Table 3
Secondary outcome measures
Vaginal insert (N = 70) Cervical gel (N = 70) Significance (two-tailed)
Changes in Bishop score at 6th hour
01 51 (72.9%) 53 (75.7%) N.S.
23 11 (15.7%) 14 (20.0%)
>3 8 (11.4%) 3 (4.3%)
Changes in Bishop score at 12th hour
01 20 (29.0%) 26 (37.7%) N.S.
23 30 (43.5%) 20 (29.0%)
>3 19 (27.5%) 23 (33.3%)
Vaginal delivery within 12 h 10 (18.5%) 7 (15,2%) N.S.
Vaginal delivery within 24 h 31 (57.4%) 27 (58,7%) N.S.
Delivery within 12 h 11 (15.7%) 10 (14.3%) N.S.
Delivery within 24 h 38 (54.7%) 40 (57.1%)
Delivery after >48 h 5 (7.1%) 6 (8.5%)
Hospital stay >4 days 15 (21.4%) 27 (38.6%) 0.027
Use of oxytocin 17 (24.3%) 29 (41.4%) 0.048
192 F. Facchinetti et al. / European Journal of Obstetrics & Gynecology and Reproductive Biology 119 (2005) 189193
vaginal insert group and three in the cervical gel group
reported infectious complications, most of which were
related to endometritis.
The mean length of stay in hospital was 4.51  1.90 days
for patients in the cervical gel group and 4.06  1.30 for
those in the vaginal insert group (P = 0.098). A long stay in
hospital (more than 4 days) was less frequent among patients
with vaginal inserts (21.4% versus 38.6%; Chi-square 4.89,
P = 0.027); we calculated an OR of 0.43 (95% CI = 0.19
0.97).
4. Discussion
This trial demonstrates that there are no substantial
differences in outcome in terms of induction of labour
between dinoprostone administered in the form of intra-
cervical gel and dinoprostone given as a vaginal insert. The
classic outcome measures of preinduction treatment with
prostaglandins, i.e. rate of delivery within 12 h and 24 h,
failure of induction, changes in cervical ripening and clinical
complication rate, were quite similar with both routes of
administration. These conclusions are in keeping with those
reported in the two previous meta-analyses [3,4] on this
subject.
Interestingly, the CS rate in our population of nulliparous
women seems to be lower with use of the vaginal insert. A
similar observation was reported after two earlier studies
that included patients with very unfavourable cervical
conditions such as those in our series [5,6], whereas in four
other similar trials the abdominal delivery rate was not
influenced by the route of dinoprostone administration [7
10]. The different rates of nulliparity among the women
taking part in the above studies would explain such
differences. Our observations should in any case be
investigated further, since this study did not have sufficiently
high power and almost twice as many cases would be
necessary for reliable demonstration of a lower CS rate with
the vaginal insert than with the intracervical gel.
As already reported [11] almost half the patients given
dinoprostine for preinduction cervical ripening started
labour and then delivered with no further intervention.
According to the present trial this seems to be true
particularly for those treated with the vaginal insert, who
required less oxytocin stimulation than those treated with
cervical gel. A similar observation was reported in a pilot
study [12]. The conclusion reached by Sanchez-Ramos et al.
[3] was precisely the opposite: they found the dinoprostone
vaginal insert less effective than other prostaglandin
preparations in the induction of active labour. Their
conclusions, however, are based mainly on their own study,
which was conducted to compare the vaginal insert with
misoprostol, which induced labour in 100% of the patients
treated with it [13]. When only studies in which
dinoprostone alone was used are considered the numbers
of patients going into labour with no further intervention
after administration of the vaginal insert and after
administration of vaginal/cervical gel was the same [14].
In keeping with earlier observations, the rate of
hyperstimulation/tachysystole affecting the fetal heart rate
and thus requiring pharmacological intervention was very
low, albeit still twice as high in women receiving vaginal
inserts than in the cervical gel group. In contrast, the rate of
major infectious complications was the same in both
groups.
Interestingly, the number of days in hospital appeared to
be lower in subjects treated with vaginal inserts than in those
treated with cervical gel. This result could be explained by
the higher proportion of vaginal deliveries in the former
group, since unless there were any complications patients
were discharged on the second postpartum day after a
vaginal delivery and not until the fourth postpartum day
when their babies had been delivered by CS. This finding
could be helpful in terms of both pharmacoeconomics and
psychosocial issues. Coming home earlier would not only be
advantageous insofar as it would reduce the costs of
hospitalisation/insurance, but also because the mother
infant dyad would benefit and the family re-organisation
would be more easily and quickly achieved.
In conclusion, the more challenging cases of preinduction
of labour at term are associated with an increased frequency
of obstetric interventions, such as an increased CS rate and
the need for more oxytocin inductions. In terms of success
and failure, vaginal inserts releasing dinoprostone are no
different from dinoprostone by the traditional cervical route.
However, use of the vaginal insert does reduce the frequency
of some obstetric interventions and allows a definite
reduction in the length of hospital stay.
References
[1] Rayburn WS. Prostaglandin E2 gel for cervical ripening and induc-
tion of labor: a critical analysis. Am J Obstet Gynecol 1989;160:
52934.
[2] Maslow AS, Sweeny AL. Elective induction of labor as a risk factor for
cesarean delivery among low-risk women at term. Obstet Gynecol
2000;95:91722.
[3] Sanchez-Ramos L, Kaunitz AM, Delke I, Gaudier FL. Cervical
ripening and labor induction with a controlled-release dinoprostone
vaginal insert: a meta-analysis. Obstet Gynecol 1999;94:87883.
[4] Hughes EG, Kelly AJ, Kavanagh J. Dinoprostone vaginal insert for
cervical ripening and labor induction: a meta-analysis. Obstet Gynecol
2001;97:84755.
[5] Miller AM, Rayburn WF, Smith CV. Patterns of uterine activity after
intravaginal prostaglandin E2 during preinduction cervical ripening.
Am J Obstet Gynecol 1991;165:10069.
[6] Smith CV, Rayburn WF, Miller AM. Intravaginal prostaglandin E2 for
cervical ripening and initiation of labor. Comparison of a multidose gel
and single, controlled-release pessary. J Reprod Med 1994;39:3814.
[7] Ottinger WS, Menard MK, Brost BC. A randomized clinical trial of
prostaglandin E2 intracervical gel and a slow release vaginal pessary
for preinduction cervical ripening. Am J Obstet Gynecol 1998;179:
34953.
[8] Hennessey MH, Rayburn WF, Stewart JD, Liles EC. Pre-eclampsia
and induction of labor: a randomized comparison of prostaglandin E2
 F. Facchinetti et al. / European Journal of Obstetrics & Gynecology and Reproductive Biology 119 (2005) 189193
as an intracervical gel, with oxytocin immediately, or as a sustained-
release vaginal insert. Am J Obstet Gynecol 1998;179:12049.
[9] Wing DA, Ortiz-Omphroy G, Paul RH. A comparison of intermittent
vaginal administration of misoprostol with continuous dinoprostone
for cervical ripening and labor induction. Am J Obstet Gynecol
1997;177:6128.
[10] Green C, Pedder G, Mason G. A randomized trial of Propess against
gel for the induction of labour at term. Br J Obstet Gynecol
1998;(Suppl. 17):105.
[11] Facchinetti F, Neri I, Genazzani AR. Factors predicting labour onset in
patients treated with prostaglandin E2 for cervical ripening. Eur J
Obstet Gynecol Reprod Biol 1995;60:12932.
193
[12] Duhl A, Tolosa J, Levia M, Nemiroff R. Randomized trial of intra-
vaginal gel, intravaginal time release insert, and intracervical gel with
prostaglandin E2 for induction of labor. Am J Obstet Gynecol
1997;176:s113.
[13] Sanchez-Ramos L, Peterson DE, Delke I, Gaudier FL, Kaunitz AM.
Labor induction with prostaglandin E1 misoprostol compared with
dinoprostone vaginal insert: a randomized trial. Obstet Gynecol
1998;91:4015.
[14] Rayburn WF, Wapner RJ, Barss VA, Spitzberg E, Molina RD, Man-
dsager N, et al. An intravaginal controlled-release prostaglandin E2
pessary for cervical ripening and initiation of labor at term. Obstet
Gynecol 1992;79:3749.