Share this article on:

Randomized Trial of Intracervical Versus Posterior Fornix Dinoprostone for Induction of Labor

Perry, Michelle Y. MD; Leaphart, W Lynn MD

doi: 10.1097/01.AOG.0000109217.24211.9D
Original Research

OBJECTIVE: To investigate whether intracervical placement of a sustained-release dinoprostone insert decreased the length of time to delivery when compared with placement in the posterior fornix.

METHODS: Sixty-three patients were randomized to intracervical (n = 33) or posterior fornix (n = 30) placement of the initial dose. Dinoprostone was placed under direct visualization with a vaginal speculum and packing forceps. The patients and staff were blinded to the site of placement. Multiple end points were examined throughout labor. Student t test, Fisher exact test, Wilcoxon test, Mann–Whitney U test, and χ2 analyses were performed when appropriate. A P value of less than .05 was considered significant.

RESULTS: Forty-six patients who required only a single dose of dinoprostone had a reduced time to delivery with intracervical (n = 24; 11.70 hours) compared with intravaginal (n = 22; 16.20 hours) placement (P = .025). There was also a reduced time to active labor (intracervical = 8.25 hours, posterior fornix = 11.50 hours; P = .083), ruptured membranes (intracervical = 10.25 hours, posterior fornix = 12.00 hours; P = .047), and request for initial pain medications (intracervical = 5.00 hours, posterior fornix = 11.25 hours; P = .025) with intracervical placement. There was no difference in number of patients managed with artificial rupture of membranes. There was no difference in maternal age, race, parity, maternal height or weight, or indication for induction. There was also no difference in cesarean delivery rate, antepartum fever, hyperstimulation, Apgar scores, birth weight, or umbilical artery pH.

CONCLUSION: In patients who respond to a single sustained-release dinoprostone insert, intracervical placement decreases time to delivery without increasing the cesarean delivery rate, infectious morbidity, or other complications of labor.


Intracervical placement of a sustained-release dinoprostone insert decreases the time to delivery when compared with placement in the posterior fornix.

From the Memorial Health University Medical Center, Mercer School of Medicine, Savannah Campus, Savannah, Georgia.

Received May 16, 2003. Received in revised form September 9, 2003. Accepted September 11, 2003.

Address reprint requests to: Michelle Y. Perry, MD, Bradley-Polk OB/GYN, 55 25th Street, NW, Cleveland, TN 37311; e-mail:

From 15% to 20% of all pregnancies require induction of labor.1 Cervical ripening before induction with oxytocin is required in 50% of these pregnancies.1 It is well established that induction of labor with oxytocin in patients with a favorable cervix leads to shorter labor, less operative deliveries, and fewer cesarean deliveries.2,3 Prostaglandins (PGs) are thought to play a significant role in the process of cervical ripening and initiation of labor.2,3 Fetal membranes produce PGE2 during pregnancy and labor.1–4 Release of this hormone leads to changes in the biochemistry of the cervix and also stimulates the production of PGF2α.5 In turn, PGF2α sensitizes the myometrium to oxytocin.5 Exogenous administration of PGE2 (dinoprostone) is known to mimic this natural process and lead to cervical ripening or labor.3,6–9 Two formulations of PGE2 have been approved by the US Food and Drug Administration for cervical ripening. Intracervical gel formulations of PGE2 (ie, Prepidil, Upjohn Pharmaceuticals, Kalamazoo, MI) have been inserted with a syringe applicator. However, passive leakage of gel from the cervix resulted in a limited amount available for absorption.3,10,11 The sustained-release PGE2 vaginal insert (Cervidil, Forest Pharmaceuticals, St. Louis, MO) has the benefits of controlled-release dosing, an easy retrieval system, and a shortened time to delivery.7,10–13 We hypothesized that intracervical placement of the sustained-release dinoprostone insert would result in an increased exposure to PGE2 and improve efficacy. The objective of this study was to investigate whether intracervical placement of a sustained-release dinoprostone insert decreased the length of time to delivery when compared with placement in the posterior fornix.

Back to Top | Article Outline


The study was conducted from January 2001 through June 2002 at Memorial Health University Medical Center in Savannah, Georgia. This study was approved by the institutional review board, and informed consent was obtained from all study subjects. Patients were included if they were admitted to the labor and delivery unit for dinoprostone induction with a singleton, cephalic fetus greater than or equal to 36 weeks’ gestation with an initial Bishop score less than 614 (Table 1). Patients were excluded if they had cervical dilatation greater than 2 cm, a history of prior cesarean delivery or hysterotomy, ruptured membranes, intrauterine growth restriction, known fetal anomalies, or a nonreassuring fetal heart rate pattern. Patients with prior use of cervical ripening agents in the current pregnancy or other contraindications to vaginal delivery were also excluded. Subjects were randomized with the use of sealed, opaque envelopes. One hundred twenty envelopes contained information on the site of placement, with equal numbers of assignments to both intracervical and intravaginal placement. Envelopes were then shuffled and placed into a covered box with subsequent blind retrieval. Patient demographics, including age, race, height, weight, gravidity, parity, gestational age, and indication for induction of labor were recorded. The resident physician performed a digital cervical examination before randomization to verify the need for cervical ripening. A Bishop score was assigned at this time.

Once patients were found appropriate for the study and consents were obtained, an envelope containing group assignment was drawn from the box. The 10-mg dinoprostone insert was placed by direct visualization with a vaginal speculum and packing forceps. With the exception of the resident placing the dinoprostone, the staff and patients were blinded to the site of placement. The dinoprostone insert remained in place for 12 hours or until labor ensued, at which time augmentation with oxytocin was initiated. If at any time during the ripening process the dinoprostone was removed, subsequent doses of dinoprostone were placed in the posterior fornix by digital examination. Continuous electronic fetal monitoring and tocodynamometry were used in all participants. With removal of the dinoprostone insert, the resident performed a subsequent digital examination to determine cervical change and to assign a second Bishop score. If fetal heart rate decelerations were noted at any time during labor, assessment of the patient was performed to determine the occurrence of hyperstimulation (prolonged [greater than 2 minutes] or frequent contractions [greater than 6 per 10 minutes] with associated fetal heart rate abnormalities).15 Episodes of fetal heart rate decelerations were managed actively with oxygen therapy, maternal positioning, removal of the dinoprostone insert, discontinuation of oxytocin infusion, intravenous fluid bolus, amnioinfusion, or use of subcutaneous terbutaline when appropriate. Patients were examined with onset of pain or at any time that evidence existed for possible cervical change (ie, frequent painful contractions, spontaneous rupture of membranes, or fetal heart rate abnormalities). Active management of labor was attempted in all patients.17 The incidence of chorioamnionitis was evaluated in each patient according to parameters such as fever (temperature greater than 38C), maternal (greater than 110 beats per minute) or fetal tachycardia (greater than 170 beats per minute), uterine tenderness, or malodorous amniotic fluid. Patients who were undelivered after 12 hours of oxytocin infusion were evaluated for either repeat cervical ripening or continued oxytocin augmentation. In patients requiring further cervical ripening, subsequent doses were placed in the posterior fornix by digital examination. At the time of delivery, fetal weight, Apgar scores, and umbilical artery pH were recorded.

Based on previous research, initial statistical analysis with a two-tailed μa value of .05 and a power of .80 determined that 120 patients (60 in each arm) would be required to demonstrate a 25% reduction in time to delivery.10 Analysis was performed with Student t test, Wilcoxon test, and Mann–Whitney U test for continuous variables, and Fisher exact test and χ2 analysis for categoric variables. Statistical significance was defined as a P value less than .05. Statistical analysis was performed with SPSS 11.5 (SPSS Inc., Chicago, IL) and Internet-based statistical programs (Web χ2 calculator, available at; and Graphpad, available at This study ended before enrolling the planned sample size because the principal investigator for this resident research project graduated and left the institution.

Back to Top | Article Outline


A total of 63 patients were randomized to either intracervical (n = 33) or posterior fornix (n = 30) placement (Figure 1). The principal indications for induction of labor were oligohydramnios (intracervical = 10, posterior fornix = 5; P = .25), gestational age greater than 41 weeks (intracervical = 9, posterior fornix = 11; P = .59), pregnancy-induced hypertension or preeclampsia (intracervical = 8, posterior fornix = 7; P = 1.0), and diabetes (intracervical = 3, posterior fornix = 4; P = .70). There was no significant difference in maternal age, race, gestational age, parity, maternal height, or maternal weight (Table 2). There was no significant difference in initial Bishop scores or subsequent assigned scores (Table 3). Further analysis of the individual Bishop score components also revealed no difference. There was a decreased median time to delivery with intracervical placement, although this difference was not statistically significant (intracervical = 22.4 hours, posterior fornix = 25.1 hours; Mann–Whitney U test, P = .169). Of these patients, 46 required only a single dose of dinoprostone for cervical ripening (intracervical = 24, posterior fornix = 22; P = 1.0). These 46 patients were further analyzed. Intracervical placement resulted in a significant reduction of time to delivery, time to initial request for pain medications, and time to ruptured membranes (Table 4). There was no difference in the incidence of artificial rupture of membranes between groups (intracervical = 15, posterior fornix = 13; P = 1.0). The time to active labor or the time to onset of regular contractions was not different between groups (Table 4). Additional analysis of these 46 patients revealed no significant difference in the median duration of oxytocin infusion between groups (Table 5). Of patients in the intracervical group, 42% did not require oxytocin augmentation, compared with 27% in the posterior fornix group. The observed difference in time to delivery seems to be a result of a more rapid response to intracervical placement of dinoprostone; however, this does not quite reach statistical significance (Table 5). Evaluation of secondary outcomes revealed no difference in the number of patients in each group requiring additional doses of dinoprostone (intracervical = 9, posterior fornix = 9; P = .50), incidence of tachysystole (intracervical = 1, posterior fornix = 0; P = 1.0), hyperstimulation (intracervical = 3, posterior fornix = 3; P = 1.0), or the incidence of maternal fever (intracervical = 1, posterior fornix = 1; P = 1.0). Birth weights, Apgar scores, and umbilical artery pH were similar among groups (Table 6). There was no difference in the incidence of cesarean delivery (intracervical = 9, posterior fornix = 7; P = .76). No difference in indications for cesarean delivery was noted between groups. The primary indications included failure to progress (intracervical = 4, posterior fornix = 6; P = .15), nonreassuring fetal heart rate pattern (intracervical = 4, posterior fornix = 0; P = .09), cephalopelvic disproportion (intracervical = 0, posterior fornix = 1; P = .44), and chorioamnionitis (intracervical = 1, posterior fornix = 0; P = 1.0). Although there was a trend toward more abnormalities of fetal heart rate in the intracervical group, further analysis revealed that all four episodes occurred during oxytocin infusion. There were no cesarean deliveries performed for nonreassuring fetal heart rate patterns during the cervical ripening process.

Back to Top | Article Outline


Prostaglandins produced by fetal membranes are known to lead to biochemical changes in the cervix.1–6 Therefore, the cervix likely contains receptors to these hormones. It is well established that PGE2 for cervical ripening and induction of labor in patients at term has been found to be of benefit.6–11 Studies also show that the mean time of delivery for patients induced via intravaginal, sustained-release dinoprostone is approximately 22 hours.6,13 Studies have been performed comparing an intracervical dose of a shorter-acting dinoprostone gel with the intravaginal sustained-release insert, revealing a decrease in the time to delivery with the sustained-release dinoprostone insert.9,10 Until now, intracervical sustained-release dinoprostone has not been tested against intravaginal placement. An exhaustive literature search with Medscape and PubMed (1982 to September 2003; all languages; search terms: “intracervical prostaglandin E2,” “intracervical dinoprostone,” “dinoprostone insert,” “sustained/controlled release dinoprostone,” “Cervidil,” “prostaglandin E2”) failed to locate any published studies related to intracervical use of the sustained-release dinoprostone insert. Our study suggests that intracervical placement of the sustained-release dinoprostone insert decreases time to delivery by 4.5 hours. There does not seem to be an increased incidence of tachysystole, hyperstimulation, or chorioamnionitis when compared with intravaginal placement (although evaluation of these outcomes is limited owing to the infrequent occurrence in this study). This study lacks sufficient power to detect differences in these variables if a difference does in fact exist. Insufficient power might also explain why a difference in duration of dinoprostone administration or oxytocin infusion was not detected. Subsequent studies with greater enrollment are required to evaluate these variables. The potential for bias could have been introduced because the physician placing the insert was knowledgeable of its placement and often performed the second Bishop score. However, there was no significant difference in Bishop scores between the two groups. Measurement bias should have been minimized by active management of labor and by having different physicians care for each patient throughout the labor course. Intracervical placement of dinoprostone contributed to a decreased time to delivery when compared with placement in the posterior fornix. There was also a reduction in time to rupture of membranes and time to request for pain medications. In this study, the decreased time to delivery is attributed to a decreased time required for cervical ripening with intracervical placement of dinoprostone, without increased risk to the patient or fetus.

Back to Top | Article Outline


1.American College of Obstetricians and Gynecologists. Induction of labor. ACOG technical bulletin no. 217. Washington: American College of Obstetricians and Gynecologists, 1995.
2.Cunningham FG, Gant NF, Leveno KJ, Gilstrap LC III, Hantu JC, Wenstrom KD. Normal labor and delivery: parturition. In: Cunningham FG, editor. Williams obstetries. 21st ed. New York (NY): McGraw-Hill; 2001. p. 280–3.
3.Rayburn WF. Prostaglandin E2 gel for cervical ripening and induction of labor, a critical analysis. Am J Obstet Gynecol 1989;160:529–34.
4.Leppart PC. Anatomy and physiology of cervical ripening. Clin Obstet Gynecol 1995;38:267–79.
5.Miller AM, Rayburn WF, Smith CV. Patterns of uterine activity after intravaginal prostaglandin E2 during preinduction cervical ripening. Am J Obstet Gynecol 1991;165:1006–9.
6.Chyu JK, Strassner HT. Prostaglandin E2 for cervical ripening: A randomized comparison of Cervidil versus Prepidil. Am J Obstet Gynecol 1997;177:606–11.
7.Miller AM, Lorkovic M. Prostaglandin E2 for cervical ripening. MCN Am J Matern Child Nurs 1993;23–30.
8.Rayburn WF, Wapner RJ, Barss VA, Spitzberg E, Molina RD, Mandsager N, et al. An intravaginal controlled-release prostaglandin E2 pessary for cervical ripening and initiation of labor at term. Am J Obstet Gynecol 1992;79:374–9.
9.Hales KA, Rayburn WF, Turnbull GL, Christenson HD, Patatanian E. Double-blind comparison of intracervical and intravaginal prostaglandin E2 for cervical ripening and induction of labor. Am J Obstet Gynecol 1997;177:606–11.
10.Stempel JE, Prins RP, Dean S. Preinduction cervical ripening: A randomized prospective comparison of the efficacy and safety of intravaginal and intracervical prostaglandin E2 gel. Am J Obstet Gynecol 1997;176:1305–9.
11.Forest Pharmaceuticals. Cervidil package insert. St. Louis: Forest Pharmaceuticals, 1995.
12.Rabl M, Elmar A, Yuksel Y, Egarter C. A randomized trial of vaginal prostaglandin E2 for induction of labor. J Reprod Med 2002;47:115–9.
13.Sanchez-Ramos L, Kaunitz AM, Delke I, Gaudier FL. Cervical ripening and labor induction with a controlled-release dinoprostone vaginal insert: A meta-analysis. Am J Obstet Gynecol 1999;94:878–83.
14.Bishop EH. Pelvic scoring for elective induction. Obstet Gynecol 1964;24:266–8.
15.Curtis P, Evans S, Resnick J. Uterine hyperstimulation: The need for standard terminology. J Reprod Med 1987;32:91–5.
16.O’Driscoll K, Meagher D. Active management of labour. London: Sanders, 1980.
© 2004 The American College of Obstetricians and Gynecologists