Induction of labor is commonly performed at term for a variety of maternal and fetal indications. The rate of labor induction has more than doubled from 9.0% in 1989 to 20.6% in 2002.1 The increase in the frequency of labor induction is likely in part due to practitioners performing elective induction for marginal indications or without a clear medical indication.2 Physicians are often pressured to “do something” after a women's due date has arrived, and it is often difficult not to intervene. Elective induction is not without consequences, such as increased cesarean delivery rates in nulliparous women and the potential for increased health care cost.3,4
A few studies have investigated alternatives to induction of labor and management of the nonripe cervix at term. One such alternative is outpatient cervical ripening without subsequent induction. Outpatient cervical ripening with prostaglandin E2 has been shown to be effective in improving Bishop scores and decreasing the length of gestation.5–7 Stitely et al8 found one to two applications of intravaginal prostaglandin E1 (misoprostol) at 41 weeks gestation decreased the length of gestation. Incerpi et al9 used a similar protocol and found 1 to 2 serial applications of intravaginal misoprostol were effective in reducing the need for induction in women with gestational diabetes at 38.5 weeks. In the current study we sought to estimate the effect on the interval to delivery of a single outpatient intravaginal dose of misoprostol in women with gestations beyond 40 weeks.
MATERIALS AND METHODS
This was a randomized, double blind, placebo-controlled trial. The Institutional Human Subjects Committee at Wright-Patterson U.S. Air Force Base Medical Center approved the protocol in November of 2000. Women older than 18 years of age who had completed at least 40 weeks of pregnancy, with a Bishop score of less than 9, and a well-dated pregnancy were eligible for inclusion. Obstetric dating was per the American College of Obstetricians and Gynecologists criteria for confirmation of a term gestation.10 Exclusion criteria consisted of any current obstetric indication for induction of labor including oligohydramnios defined as an amniotic fluid index of less than 5.0 cm, ruptured membranes, the presence of any contraindication to a trial of labor including fetal malpresentation or vaginal bleeding, multiple gestation, regular uterine contractions defined as 10 or more per hour, and a history of a prior cesarean delivery. Women were identified for the study by the resident and staff physicians who were providing their prenatal care.
Each participant was assigned a sequential study number, and randomization was performed by the hospital's inpatient pharmacy using a computer-generated random number table. Women were randomly assigned to receive a one-time outpatient intravaginal administration of 25 μg of misoprostol or placebo consisting of an inert substance. The randomization list was kept in the pharmacy and the pharmacists had no other involvement in the study. The investigators and study participants were blinded to group assignments.
Misoprostol was prepared by crushing 100 μg tablets (Cytotec, G.D. Searle, Chicago, IL) into powder, dividing it by weight into 25-μg aliquots, and inserting it into opaque gel capsules. Placebo was comprised of starch (United States Pharmacopeia), an inert substance, enclosed in identical gel capsules. The study medications were then placed in identical individual amber-colored containers. The medication was stored in the hospital pharmacy at room temperature in a standard humidity-controlled environment. Additional study medication was prepared every 2 to 3 months as necessary. Informed consent was obtained from each participant after a careful explanation of the study, and the next randomly assigned sequential study medication was brought to the obstetric clinic from the pharmacy.
One of the investigators determined fetal presentation, and a nonstress test (NST) and cervical examination were performed. A Bishop score was assigned.11 If variable decelerations were noted on the NST, an amniotic fluid index was determined. The gel capsule was placed in the posterior vaginal fornix, adjacent to the cervix. Fetal and uterine monitoring was continued for one hour. If there was no evidence of uterine hyperstimulation or tachysystole, the woman was discharged home with labor precautions and instructed to follow up with her regularly scheduled prenatal care. The women's clinic physicians provided the remainder of the prenatal care. The physicians were aware of their patients’ study participation, although they did not know the group assignment. The women's physicians made all obstetric decisions without input from the investigators. The study participants’ physicians made all decisions regarding induction of labor and all inductions were for standard obstetric indications.10 Labor was routinely induced if the Bishop score was greater than 8 after 41 weeks gestation, and after 42 weeks with inpatient cervical ripening if necessary. The latter are referred to as “postterm” inductions. Participants who presented with premature rupture of membranes were administered oxytocin to stimulate labor, but in the analyses were not considered to have undergone induction of labor. Admission to labor and delivery was reserved strictly for women in active labor or those with premature rupture of membranes.
Active labor was defined as progressive cervical dilation beyond 3 cm in the presence of regular uterine contractions. The length of labor was defined as the time of the onset of active labor to completion of the second stage of labor. Febrile morbidity was defined as maternal clinical chorioamnionitis or endomyometritis, defined by standard characteristics.12 The change in hemoglobin was calculated from the difference in hemoglobin measured at admission to labor and delivery and hemoglobin measured at 12–24 hours after delivery. Delivery and maternal outcome data were obtained from computerized labor records. Maternal demographics, admission, delivery, and neonatal outcome data were recorded.
All subjects were analyzed by intention to treat. Statistical and data analyses were performed with JMP Statistical Discovery Software (SAS Institute Inc., Cary, NC), GraphPad Prism 4.0 (GraphPad Software, San Diego CA), and Microsoft Excel (Microsoft Corporation, Redmond, WA). An α value of 0.05 was considered significant. The primary outcome, interval to delivery, was defined as the time interval from gel capsule placement to delivery. The null hypothesis was that there would be no difference in the mean interval to delivery between the 2 groups. We assumed a difference in the mean interval to delivery of 7.0 ± 8.0 days would be clinically significant. Power analysis (α of 0.05 and a β of 0.2) determined 44 subjects (22 per group) were necessary to achieve statistical significance. The distribution of continuous variables was tested for normality by the Kolmogorov-Smirnov test. Student t test and analysis of variance were used to compare variables with normal distributions. The Mann-Whitney test was used to compare nonparametric continuous variables. The Fisher exact test was used for comparisons involving 2 groups of nominal or ordinal variables. Survival curves were compared by the log-rank test.
The study took place from March 2001 to March 2003. Sixty-eight women consented to participate. Thirty-three subjects were randomly assigned to receive misoprostol, and 35 were assigned to receive placebo. Three subjects in the placebo group and 1 in the misoprostol group were excluded after randomization and did not receive the study medication. Two had regular contractions every 2 to 4 minutes, and 2 were found to have oligohydramnios after variable decelerations were noted on the NST. The 4 were analyzed in their respective groups (intent to treat).
Table 1 contains the maternal demographics. There were no significant differences between the groups. The mean Bishop score ± standard deviation and (95% confidence interval) at the time of placement was 4.7 ± 1.7 (4.1–5.3) for the misoprostol group compared with 4.3 ± 2.2 (3.5–5.1) for the placebo group, P = .4. The mean Bishop score at the time of admission to labor and delivery was significantly greater in the misoprostol group at 9.0 ± 2.8 (8.0–10.0) compared with 7.5 ± 2.8 (6.5–8.5) for the placebo group, P = .03.
All but 1 study subject were discharged home after 1 hour of monitoring. There was 1 case of uterine hyperstimulation in a nulliparous woman who had received misoprostol. She had late decelerations with irregular uterine contractions after the medication placement. She was taken to labor and delivery and subsequently had uterine contractions every 2 minutes and persistent late decelerations. Her cervix was dilated 2 cm, and a primary cesarean delivery was performed. The umbilical arterial blood pH was 7.15, and there was no neonatal or maternal morbidity. Five women, all in the misoprostol group, returned and were admitted to labor and delivery less than 6 hours after placement of the study medication (range 2.9–4.8 hours).
The delivery outcomes are summarized in Table 2. The interval to delivery for both groups passed the Kolmogorov-Smirnov test for normality. The mean interval to delivery was significantly less for the misoprostol group than the placebo group, 4.2 ± 4.1 (2.8–5.7) days compared with 6.1 ± 3.6 (4.9–7.3) days, P = .04. The difference was greater when only the nulliparous women were compared, 4.2 ± 4.0 (2.8–5.6) days for the misoprostol group (n = 20) compared with 7.2 ± 3.7 (5.9–8.5) days for the placebo group (n = 19), P = .02. Within 48 hours of medication placement, 15 of 33 (45%) of women in the misoprostol group had delivered compared with 5 of 35 (14%) in the placebo group, P = .007; within 7 days 23 of 33 (70%) in the misoprostol group had delivered compared with 19 of 35 (54%) in the placebo group, P = .2. Figure 1 contains the survival curve analysis for interval to delivery in days, censored for subjects that underwent induction of labor. The curves are significantly different, P = .04. The median interval to delivery for the misoprostol group was 4.1 days compared with 9.2 days for the placebo group.
There were 12 of 33 (36%) inductions of labor in the misoprostol group compared with 19 of 35 (54%) in the placebo group, P = .15. The indications for induction consisted of postterm (5 misoprostol, 10 placebo), hypertensive disorder (4 misoprostol, 3 placebo), oligohydramnios (1 misoprostol, 4 placebo), abnormal fetal heart rate (2 misoprostol), and Bishop score greater than 8 after 41 weeks (2 placebo). There were 10 inductions of labor in the nulliparous women in both the misoprostol (50%) and placebo (53%) groups, P = 1.0. The cesarean delivery rate was 27% (9/33) in the misoprostol group compared with 26% (9/35) in the placebo group, P = 1.0. The cesarean delivery rate was not significantly different between the groups for nulliparous patients, 8 of 20 (40%) misoprostol compared with 7 of 19 (37%) placebo, P = 1.0. The indications for cesarean delivery consisted of arrest of labor (5 in the misoprostol group, 7 in the placebo group), nonreassuring fetal heart tracings (3 in misoprostol, 2 in the placebo group), and breech presentation (1 in the misoprostol group).
The 2 groups did not have a significance difference in other delivery outcomes, including length of active labor, use of labor epidurals, operative vaginal deliveries, episiotomy, vaginal lacerations, and third-degree lacerations, Table 2. There were no fourth-degree lacerations in either group.
Meconium-stained amniotic fluid was present in 6 of 33 (18%) deliveries in the misoprostol group compared with 8 of 35 (23%) in the placebo group, P = .8. Two deliveries in the placebo group (5.7%) had 5-minute Apgar scores less than 7, and there were none in the misoprostol group, P = .49. There was no difference in febrile morbidity: 6% (2/33) misoprostol compared with 20% (7/35) placebo, P = 1.0; or in the mean change in hemoglobin: 2.2 ± 1.5 g/dL misoprostol compared with 1.8 ± 1.1 g/dL placebo, P = .19. There were no neonatal intensive care unit admissions in the misoprostol group and one in the placebo, P = 1.0.
Outpatient cervical ripening with a single 25-μg intravaginal dose of misoprostol in women with unfavorable cervices at term was effective in improving the Bishop score and decreasing the mean interval to delivery. Nulliparous women who received misoprostol had the greatest benefit in decreasing the mean interval to delivery. Fifty percent of the women who received misoprostol delivered within 4.1 days, which was less than half the median duration of pregnancy (9.2 days) of the placebo group. There were no significant complications in either group and no differences in low Apgar scores at 5 minutes, febrile morbidity, or prepartum and postpartum change in maternal hemoglobin.
The study lacked the power to show a significant difference in the number of inductions of labor and cesarean delivery rate; however, the placebo group did have twice the number of postterm inductions (n = 10) compared with the misoprostol group (n = 5). A larger study would be needed to determine whether a single dose of misoprostol decreases the number of inductions or cesarean deliveries. For example, 125 subjects would be required in each group to attain statistical significance (α of 0.05 and a β of 0.2) for the 18% difference we found in the induction rate. An even greater number of subjects would be required to assess the effect on the cesarean delivery rate. Misoprostol has generally been accepted as a safe and effective means of inpatient cervical ripening.10 We did not find any differences in measures of maternal or neonatal morbidity, but the study lacks the power to draw conclusions about the safety of outpatient cervical ripening with misoprostol.
This study is limited by the possibility of a type II error: the actual standard deviations and differences in the mean interval to delivery between the groups were less than the assumed values used for the power analysis (7.0 ± 8.0 days). A post hoc power analysis determined there was approximately 50% power to detect a difference in the mean interval to delivery for all patients and approximately 60% power for only the nulliparous patients. One hundred forty subjects would have to be enrolled to attain 80% power for both outcomes.
Outpatient cervical ripening with misoprostol is potentially an excellent alternative to induction of labor, especially when there are marginal indications for induction. Other investigators have shown serial dosing of outpatient vaginal misoprostol to be effective in decreasing the length of gestation.8,9 Stitely et al,8 in a blinded placebo controlled trial, compared serial doses of 25 μg of vaginal misoprostol with placebo in pregnancies at 41 weeks or greater gestation and Bishop scores of 4 or less. After the initial dose of study medication, a second dose was administered 24 hours later, and labor was induced after 48 hours if the subject was still pregnant. Eighty-nine percent of the women who received misoprostol entered active labor within 48 hours.8 Incerpi et al9 in a double-masked, placebo-controlled trial compared an initial dose of 25 μg of vaginal misoprostol with placebo in women with gestational diabetes at 38.5 weeks and Bishop scores less than 9. The misoprostol was repeated 3 days later if the Bishop score remained less than 9. Labor was induced 7 days after the initial dose if spontaneous labor did not develop. They found misoprostol to be effective in reducing the need for labor induction and the induction-delivery interval.9 This study is unique in that it is a simple protocol requiring 1 outpatient visit and administration of a single outpatient dose of misoprostol, and induction of labor was not part of the protocol. The total duration of the outpatient office visit, examination, NST, medication placement, and monitoring averaged approximately 2 hours.
Outpatient cervical ripening without subsequent induction of labor enhances the natural process of cervical ripening. There are many potential benefits of this approach, including decreasing the numbers of prolonged pregnancies, elective inductions, and cesarean deliveries, and increased success when induction is performed. This study adds to the clinical evidence on the effectiveness of outpatient ripening with vaginal misoprostol. A single 25-μg dose of intravaginal misoprostol seems to be an effective technique for management of the unripe cervix at term, resulting in a significant decrease in the interval to delivery. Our study was conducted in a clinic that was on the same floor as the labor and delivery unit and we were able to intervene promptly in the 1 case of uterine hyperstimulation. The results should not be extrapolated to other applications. We recommend that outpatient cervical ripening with misoprostol only be performed as part of a study protocol until the safety is confidently established.
1. Martin JA, Hamilton BE, Sutton PD, Ventura SJ, Menacker F, Munson ML. Births: final data for 2002. Natl Vital Stat Rep 2003;52:1–114.
2. Rayburn WF, Zhang J. Rising rates of labor induction: present concerns and future strategies. Obstet Gynecol 2002;100:164–7.
3. Seyb ST, Berka RJ, Socol ML, Dooley SL. Risk of cesarean delivery with elective induction of labor at term in nulliparous women. Obstet Gynecol 1999;94:600–7.
4. Kaufman KE, Bailit JL, Grobman W. Elective induction: an analysis of economic and health consequences. Am J Obstet Gynecol 2002;187:858–63.
5. Sawai SK, O'Brien WF, Mastrogiannis DS, Krammer J, Mastry MG, Porter GW. Patient-administered outpatient intravaginal prostaglandin E2
suppositories in post-date pregnancies: a double-blind, randomized, placebo-controlled study. Obstet Gynecol 1994;84:807–10.
6. Magann EF, Chauhan SP, Nevils BG, McNamara MF, Kinsella MJ, Morrison JC. Management of pregnancies beyond forty-one weeks’ gestation with an unfavorable cervix. Am J Obstet Gynecol 1998;178:1279–87.
7. McKenna DS, Costa SW, Samuels P. Prostaglandin E2 cervical ripening without subsequent induction of labor. Obstet Gynecol 1999;94:11–4.
8. Stitely ML, Browning J, Fowler M, Gendron RT, Gherman RB. Outpatient cervical ripening with intravaginal misoprostol. Obstet Gynecol 2000;96:684–8.
9. Incerpi MH, Fassett MJ, Kjos SL, Tran SH, Wing DA. Vaginally administered misoprostol for outpatient cervical ripening in pregnancies complicated by diabetes mellitus. Am J Obstet Gynecol 2001;185:916–9.
10. American College of Obstetricians and Gynecologists. Induction of labor. ACOG Practice Bulletin 10. Washington, DC: ACOG; 1999.
11. Bishop EH. Pelvic scoring for elective induction. Obstet Gynecol 1964;24:266–8.
12. Duff P. Maternal and perinatal infection. In: Gabbe SG, Niebyl JR, Simpson JL, editors. Obstetrics: normal and problem pregnancies. New York (NY): Churchill Livingstone; 2002;1293–345.