Abstract
BACKGROUND:
Cervical ripening is commonly needed for labor induction. Finding an optimal route of misoprostol dosing for efficacy, safety, and patient satisfaction is important and not well studied for the buccal route.
OBJECTIVE:
To compare the efficacy and safety of vaginal and buccal misoprostol for women undergoing labor induction at term.
STUDY DESIGN:
The IMPROVE trial was an institutional review boarde-approved, triple-masked, placebo-controlled randomized noninferiority trial for women undergoing labor induction at term with a Bishop score ≤6. Enrolled women received 25 mcg (first dose), then 50 mcg (subsequent doses) of misoprostol by assigned route (vaginal or buccal) and a matching placebo tablet by the opposite route. The primary outcomes were time to delivery and the rate of cesarean delivery performed urgently for fetal nonreassurance. A sample size of 300 was planned to test the noninferiority hypothesis.
RESULTS:
The trial enrolled 319 women, with 300 available for analysis, 152 in the vaginal misoprostol group and 148 in the buccal. Groups had similar baseline characteristics. We were unable to demonstrate noninferiority. The time to vaginal delivery was lower for the vaginal misoprostol group (median [95% confidence interval] in hours: vaginal: 20.1 [18.2, 22.8] vs buccal: 28.1 [24.1, 31.4], log-rank test P = .006, Pnoninferiority = .663). The rate of cesarean deliveries for nonreassuring fetal status was 3.3% for the vaginal misoprostol group and 9.5% for the buccal misoprostol group (P = .033). The rate of vaginal delivery in <24 hours was higher in the vaginal group (58.6% vs 39.2%, P = .001).
CONCLUSION:
We were unable to demonstrate noninferiority. In leading to a higher rate of vaginal deliveries, more rapid vaginal delivery, and fewer cesareans for fetal issues, vaginal misoprostol may be superior to buccal misoprostol for cervical ripening at term.
Keywords: Buccal, cervical ripening, labor induction, misoprostol, term pregnancy, vaginal
The rate of labor induction has doubled over the last 25 years, with nearly 25% of gravid women undergoing labor induction in the United States.1 The goal of induction of labor is to achieve vaginal delivery by stimulating uterine contractions, cervical dilation, and active labor. Labor is typically induced when the risks of continuing the pregnancy outweigh the risks of delivery or labor induction.2
Often when labor is induced, the cervix must be “ripened,” a process involving cervical softening, thinning, and dilation to help facilitate the successful induction of labor.2 Prostaglan-dins have been effectively utilized for cervical ripening and labor induction for decades, as they both induce cervical changes and stimulate uterine contractions.3 Misoprostol, a synthetic prostaglandin E1 analogue, has been shown to be an effective and safe drug for induction and is the most frequently used induction method.2,4,5 According to both the World Health Organization and the ACOG Practice Bulletin on Induction of Labor, doses of 25 mcg and higher are typically administered every 4–6 hours, depending on provider preference.2
Misoprostol is currently administered in many different ways.6 It can be administered vaginally, rectally, orally, buccally, and sublingually.7,8 Although vaginal administration of misoprostol is most common, recent trends in practice have shifted toward more buccal use of this drug. A recent survey found that midwives indicated a preference toward buccal dosing.9 However, only 1 published trial directly compared buccal misoprostol (BM) with vaginal misoprostol (VM).10 In that trial of 157 women, there were no significant differences in any of the outcomes other than higher rates of tachysystole in the buccal group. However, that trial utilized higher doses of misoprostol (up to 100 mcg) than are typically used clinically.2 A recent retrospective cohort study of 207 women also found that time to delivery may be similar for the 2 routes of administration.11 Clinical experience at our center and these retrospective observations led to a hypothesis that the 2 routes were clinically equivalent. Given that if buccal was not inferior clinically to vaginal dosing and that patients may prefer avoiding additional vaginal examinations to place medication, a noninferiority trial was warranted.
The primary objective of the Induction with MisoPRostol: Oral mucosa versus Vaginal Epithelium (IMPROVE) study was to compare the efficacy and safety of VM and BM for women undergoing labor induction at greater than or equal to 37 completed weeks’ gestation with a live fetus.
Materials and Methods
The IMPROVE trial was a triple-blinded, placebo-controlled study conducted from August 2015 through October 2017 at 2 hospitals in Indianapolis, Indiana, served by the Indiana University (IU) School of Medicine Department of OB/GYN and the physicians and certified nurse midwives of the HealthNet medical group. The first participant was enrolled On October 4, 2015 and the final participant was enrolled October 3, 2017. This study was a triple-masked, randomized controlled trial for inpatient women on labor and delivery receiving misoprostol for cervical ripening during labor induction. Funding for the study was provided by the IU Department of OB/GYN. The trial was conducted under a U.S. Food and Drug Administration Investigational New Drug application (IND#: 122727), was approved by the IU institutional review board, and was registered on clinicaltrials.gov (NCT 02408315). A Data & Safety Monitoring Board reviewed blinded results for efficacy and safety after recruitment of 50 women, after recruitment of 150 women, and at trial completion. The full details of the methods of the study are presented as Supplementary Material (Appendix).
Participants
Women who presented to the labor and delivery unit for delivery who required cervical ripening were eligible for the trial. Women ≥14 years of age undergoing either a medically indicated induction of labor at a gestational age beyond 37 0/7 weeks or an elective induction of labor after 39 weeks with a singleton pregnancy in the cephalic presentation, and a modified Bishop score ≤6 (commonly used as a cutoff for the need for cervical ripening), were eligible for enrollment. Women were excluded if they had a known prior uterine scar, untreated cervical infection, known major fetal congenital anomaly, or evidence of fetal compromise (category 2 or 3 fetal tracing) before the start of the induction. All women provided informed consent.
Study drugs and preparation
Misoprostol tablets (100 mcg) were obtained from the manufacturer (Novel Laboratories, Somerset, NJ). Identical placebo tablets were obtained from University of Iowa Pharmaceuticals. Tablets were divided in half or in quarters (25- or 50-mcg doses) by the investigational pharmacies and packaged in identical foil packets labeled either “Vaginal” or “Buccal.”
Randomization and allocation concealment
Computer-generated, stratified randomization with blocks of size 10 was used for each hospital with 1:1 assignment to treatment group. After informed consent was obtained by the study team, the appropriate investigational pharmacy was notified. The pharmacist on duty obtained the next sequentially numbered study drug packet and sent it to labor and delivery. Other than the investigational pharmacist, who did not have direct subject contact, no investigators, providers, or patients had knowledge of randomization assignment.
Study procedures
After informed consent was obtained and study drugs in air-tight foil packs were acquired from the pharmacy, the tablet marked “Buccal” was placed between the teeth and mucous membrane of the cheek and the tablet marked “Vaginal” was placed by the clinical care provider high into the posterior vaginal fornix. The initial dose of misoprostol used in this protocol was 25 mcg. Subsequent doses, if utilized, were 50 mcg, in accordance with the ACOG Practice Bulletin.2 Throughout the cervical ripening and induction process, continuous external electronic fetal monitoring was utilized as per standard hospital practice.
Cervical examinations were performed approximately every 4 hours, prior to buccal and vaginal administration of the next dose. An additional dose (50 mcg) of study drug was given if clinically indicated. Study participation and drug placement continued until 1 of the following occurred: (1) there was adequate response and cervical ripening was no longer needed; (2) there were signs of tachysystole, nonreassuring fetal heart tracing, or other adverse event that would make the provider stop the misoprostol; or (3) 24 hours of study drug had been given (maximum of 7 doses). After cervical ripening was complete or the participant was taken off of the study, there were no limitations placed on the clinical care. At least 30 days after delivery, data were abstracted from the medical record to capture all relevant maternal and newborn outcomes and complications.
Outcomes assessment
The primary efficacy outcome was the time to delivery, defined as the time from placement of the first dose of study drug to the time of delivery. The primary safety outcome was the rate of cesarean delivery performed urgently for fetal nonreassurance as the primary indication; however, the study was not powered for this outcome. Secondary efficacy and safety outcomes included typical outcomes during labor and delivery, such as labor characteristics and other medications used (for full list, see Supplementary Material). The IMPROVE trial also assessed participant satisfaction with a modified tool from Nassar et al.12
At least 30 days after delivery, the medical records of the participant and her newborn were reviewed and data abstracted from the medical record to capture all relevant maternal and newborn outcomes and complications.
Statistical analysis and sample size calculations
The study was planned as a noninferiority trial to assess the primary outcome of time to delivery. This was based on our retrospective study finding similar times to delivery.11 Our data found that the median time to delivery with both BM and VM was about 18 hours. If the buccal route was noninferior, we hypothesized that participant discomfort would be less with BM, as additional vaginal examinations might be avoided. A sample size of 300 women with 260 expected vaginal deliveries11 was estimated to have 80% power to test a null hypothesis that the hazard ratio (HR) of BM relative to VM would be ≤0.74 vs alternative hypothesis that the HR would be >0.74 with type I error set at 0.05. This noninferiority margin was derived from retrospective data on time to delivery by both routes, which equated to approximately a 4.5-hour difference in median time to delivery. For all other outcomes, the 2-sided superiority P values are provided.
Participant and delivery characteristics were compared between treatment groups using appropriate tests (t test, Wilcoxon rank sum test, χ2 test, or Fisher exact test). For the primary outcome of time to delivery, median time to delivery and associated 95% confidence intervals (CI) were estimated by Kaplan-Meier method for the overall cohort and by route of delivery. The time to vaginal delivery for women who required cesarean delivery was censored at the time of cesarean delivery. Cox proportional hazards regression was used to estimate the HR for BM relative to VM for delivering vaginally and associated 95% CI. A P value < .05 provides evidence to reject inferiority and conclude the buccal route of dosing is noninferior to the vaginal route. In secondary analyses, the HR and associated 95% CI for route of misoprostol was also estimated from the Cox proportional hazards regression, adjusting for covariates known to be associated with time to delivery. All analyses were based on assigned group and completed with SAS software version 9.4 (SAS Institute, Inc, Cary, NC).
Results
Participant characteristics
Of 329 women consented for the trial (Figure 1), 10 were excluded before randomization, either because they withdrew their consent after signing or were found to not meet inclusion criteria on a secondary chart review in preparation for randomization. Thirteen randomized women (4%) did not receive the allocated intervention, most (8) because they developed an exclusion criterion before study drug placement, typically either fetal tracing abnormalities or cervical dilation change. Six women (2%) randomized and initially dosed were administratively withdrawn from analysis after later discovery of documentation or consent issues making them ineligible. This left 300 women (94% of those randomized) eligible for analysis, of whom 152 received VM and 148 received BM (Figure 1).
FIGURE 1.
CONSORT diagram of participant flow through IMPROVE trial
Demographic characteristics by treatment group are provided in Table 1. Overall, randomization achieved balanced groups. The most common indications for induction were late-term pregnancy (27%), hypertensive disorder (21%), and diabetes mellitus (10%). Hispanic women comprised 29% of the cohort and 31% of women were African American. Forty-one percent of the women were nulliparous. The mean gestational age of both groups was just over 39 weeks. A total of 242 women (80.7%) had a successful induction and delivered vaginally (Table 2). One woman in the trial was sent home after receiving a course of study drug, returned at a later date, and delivered vaginally; this participant was censored at time of discharge for the primary outcome of time to delivery but was included in all other analyses. Reasons for stopping misoprostol use were typical for cervical ripening and included sufficient cervical ripening (34%); safety concerns, most commonly tachysystole or fetal heart tracing abnormalities (27%); active labor (6%); or multiple reasons together (23%) (Table 2).
TABLE 1.
Demographic characteristics of women in the IMPROVE study at study entry
| Characteristic | VM group (n = 152) | BM group (n = 148) |
|---|---|---|
| Age (years), mean ± SD | 28.2 ± 6.4 | 27.6 ± 6.4 |
| Site, n (%) | ||
| Eskenazi | 86 (56.6) | 83 (56.1) |
| Methodist | 66 (43.4) | 65 (43.9) |
| Nulliparous women, n (%) | 59 (38.8) | 65 (43.9) |
| Race/ethnicity, n (%) | ||
| White | 74 (48.7) | 66 (44.6) |
| African American | 45 (29.6) | 49 (33.1) |
| Other | 33 (21.7) | 33 (22.3) |
| Ethnicity, n (%) | ||
| Hispanic/Latino | 42 (27.6) | 44 (29.7) |
| Non-Hispanic/non-Latino | 109 (71.7) | 99 (66.9) |
| Prefer not to answer | 1 (0.7) | 5 (3.4) |
| BMI (kg/m2), mean ± SD | 35.7 ± 7.2 | 35.1 ± 7.3 |
| BMI category, n (%) | ||
| <18 | 0 | 1 (0.7) |
| 18–25 | 7 (4.6) | 12 (8.1) |
| 25–30 | 22 (14.5) | 22 (14.9) |
| 30–40 | 87 (57.2) | 77 (52.0) |
| >40 | 36 (23.7) | 36 (24.3) |
| Gestational age at trial entry (weeks), mean ± SD | 39.6 ± 1.3 | 39.5 ± 1.3 |
| Indication for induction, n (%) | ||
| Fetal indications | 10 (6.6) | 17 (11.5) |
| Hypertensive disorder | 33 (21.7) | 31 (21.0) |
| Diabetes mellitus | 10 (6.6) | 20 (13.5) |
| Late-term pregnancy (≥41+0 weeks) | 44 (29.0) | 36 (24.3) |
| Elective | 21 (13.8) | 13 (8.8) |
| Multiple reasons | 13 (8.6) | 13 (8.8) |
| Other | 21 (13.8) | 18 (12.2) |
| Epidural, n (%) | 127 (83.6) | 122 (82.4) |
| Cervical dilation, n (%) | ||
| <1 cm | 31 (20.7) | 37 (25.2) |
| 1–2 cm | 102 (68) | 89 (60.6) |
| >2 cm | 17 (11.3) | 21 (14.3) |
| Cervical dilation (cm), mean ± SD | 1.3 ± 0.9 | 1.3 ± 0.9 |
| Effacement of cervix, n (%) | ||
| 0–30% or >4 cm length | 82 (54.7) | 82 (55.8) |
| 40–50% or 3–4 cm length | 62 (41.3) | 52 (35.4) |
| 60–70% or 1–2 cm length | 5 (3.3) | 10 (6.8) |
| Effacement of cervix measure (%), mean ± SD | 31.7 ± 20.8 | 31.8 ± 22.8 |
| Fetal station (cm), n (%) | ||
| −3 | 106 (72.1) | 106 (72.6) |
| −2 | 34 (23.1) | 36 (24.7) |
| −1,0 | 7 (4.8) | 4 (2.7) |
| Fetal station (cm), mean ± SD | −2.8 ± 0.8 | −2.8 ± 0.7 |
| Bishop score, mean ± SD | 2.3 ± 1.7 | 2.2 ± 1.7 |
No statistically significant differences in any baseline characteristics between groups.
BM, buccal misoprostol; BMI, body mass index; VM, vaginal misoprostol.
TABLE 2.
Labor and delivery outcomes of women in IMPROVE study
| N = 152, 24 censored | N = 148, 34 censored | Pvaluea | HR [95% CI]a | Pvaluenoninferiority | |
|---|---|---|---|---|---|
| Time to delivery (hours), median [95% CI] | 20.1 [18.2, 22.8] | 28.1 [24.1, 31.4] | .006 | 0.70 [0.54, 0.90] | .663 |
| VM group (n = 152) | BM group (n = 148) | Pvalueb | |
|---|---|---|---|
| Cesarean for fetal nonreassurance, n (%) | 5 (3.3) | 14 (9.5) | .033 |
| Vaginal delivery in less than 24 hours, n (%) | 89 (58.6) | 58 (39.2) | .001 |
| Reason for stopping misoprostol, n (%) | .006d | ||
| Onset of active labor | 7 (4.6) | 10 (6.8) | |
| Sufficient cervical ripening | 60 (39.5) | 43 (29.1) | |
| Safety concerns | 47 (30.9) | 34 (23) | |
| Multiple reasons | 31 (20.4) | 38 (25.7) | |
| Other | 7 (4.6) | 23 (15.5) | |
| Route of delivery, n (%) | .151 | ||
| Vaginal delivery | 127c (84.2) | 115 (77.7) | |
| Cesarean delivery | 24 (15.8) | 33 (22.3) | |
| Cesarean (indications), n (%) | N = 24 | N = 33 | .376d |
| Fetal nonreassurance | 5 (20.8) | 14 (42.4) | |
| Arrest of dilation | 3 (12.5) | 4 (12.1) | |
| Arrest of descent | 3 (12.5) | 1 (3.0) | |
| Multiple reasons | 11 (45.8) | 11 (33.3) | |
| Other | 2 (8.3) | 3 (9.1) | |
| Chorioamnionitis, n (%) | 7 (4.6) | 10 (6.8) | .420d |
| Postpartum hemorrhage, n (%) | 8 (5.3) | 6 (4.0) | .620d |
| Blood transfusion, n (%) | 1 (0.6) | 1 (0.7) | .794d |
| Oxytocin use, n (%) | 100 (65.8) | 111 (75) | .081 |
| Doses of misoprostol needed to get into active labor (n), median (range) | 2.0 (1.0–5.0) | 3.0 (1.0–7.0) | <.001d |
| Maximum units of oxytocin administered, median (range) | 4.0 (0–36.0) | 6.0 (0–30.0) | .001d |
P value for noninferiority hypothesis based on Cox proportional hazards model (H0: HR ≤ 0.74 vs HA: HR > 0.74); P value < .05 provides evidence to reject inferiority and conclude BM is noninferior to VM.
BM, buccal misoprostol; CI, confidence interval; HR, hazard ratio; VM, vaginal misoprostol.
P value from log-rank tests; profile likelihood confidence intervals reported for 95% CI of HR;
P values obtained from t test (continuous) or χ2 test (categorical);
One subject was admitted into the study, got discharged, and then returned 2 weeks later to have a vaginal delivery. Delivery characteristics (except for route of delivery) are included for this person as they were censored for time to delivery;
P values obtained from Wilcoxon rank sum test (continuous) or Fisher exact test (categorical).
Primary outcomes
The primary outcome of time to delivery from receipt of first dose of misoprostol was significantly longer for women receiving BM than VM (median [95% CI]; BM: 28.1 [24.1 to 31.4] hours vs VM: 20.1 [18.2 to 22.8] hours, P = .006) (Table 2, Figure 2, top). Based on the Cox proportional hazards model, the BM vs VM HR for vaginal delivery is 0.70 (0.54 to 0.90); thus our null hypothesis that the HR ≤0.74 is not rejected (Pnoninferiority = .663) and we cannot conclude BM is noninferior to VM.
FIGURE 2. Kaplan-Meier curves illustrating time to delivery.
Top panel shows time to delivery from start of induction (in hours) for all women in IMPROVE trial, stratified by route of misoprostol. Bottom panel shows time to delivery for participants stratified by nulliparous (yes or no) and route of misoprostol. Women delivered by cesarean delivery are censored at the time of that delivery. Numbers below the graphs are the number of women in each group still pregnant at that time point.
BM, buccal misoprostol; VM, vaginal misoprostol.
For the primary safety outcome, there was higher rate of urgent cesarean delivery for fetal nonreassurance in women receiving BM (n =14, 9.5%) vs VM (n = 5, 3.3%), P = .033 (Table 2, Table S1).
Covariates known to be associated with time to delivery were included in a full Cox proportional hazards model to estimate the adjusted HR for BM vs VM for time to delivery (Table 3). In the adjusted model, the BM vs VM adjusted HR is 0.59 (0.45 to 0.77), P < .0001 (Table 3). Thus, VM was found to be superior after adjustment for covariates. Although the assumption of proportional hazards was not rejected for this full model (P = .062), the 2-way inter-action between parity and log of time to delivery and oxytocin and log of time delivery were both significant at the 0.05 level. Thus, we ran additional models to examine these important covariates (Supplementary Material, Tables S2–S4). Multiparous women had a larger difference in median time to delivery between treatment groups compared with nulliparous women (Tables S3 and S4, and Figure 2, bottom).
TABLE 3.
Multiple Cox proportional hazards regression model for time to delivery (oxytocin included as baseline covariate)
| Covariate | Estimate | SE | Pvalue | HR [95% CI]a | Pvaluenoninferiority |
|---|---|---|---|---|---|
| Dose route (BM vs VM) | −0.530 | 0.137 | .0001 | 0.59 [0.45, 0.77] | .952 |
| Site (Eskenazi vs Methodist) | 0.605 | 0.149 | <.0001 | 1.83 [1.37, 2.46] | |
| Maternal age (y) | 0.0003 | 0.012 | .982 | 1.00 [0.98, 1.02] | |
| BMI (kg/m2)b | −0.010 | 0.010 | .294 | 0.99 [0.97, 1.01] | |
| Nulliparous (no) | 1.17 | 0.171 | <.0001 | 3.04 [2.19, 4.26] | |
| Bishop score | 0.158 | 0.038 | <.0001 | 1.17 [1.09, 1.26] | |
| Epidural (no vs yes) | 0.622 | 0.168 | .0002 | 1.86 [1.33, 2.57] | |
| Need for oxytocin (no vs yes) | 1.324 | 0.154 | <.0001 | 3.76 [2.77, 5.08] |
N = 300, 242 vaginal deliveries, model Akaike information criterion = 2190.2.
HRs <1.00 should be interpreted as that group needing more time to delivery.
BM, buccal misoprostol; BMI, body mass index; CI, confidence interval; HR, hazard ratio; SE, standard error; VM, vaginal misoprostol.
Profile likelihood confidence intervals;
One person missing BMI, imputed with mean of cohort.
Secondary outcomes
The rate of vaginal delivery within 24 hours from start of induction was significantly higher in the VM group (58.6% vs 39.2%, relative risk 1.49 [1.17 to 1.90], P = .001). The number of doses of misoprostol required to achieve active labor was significantly less in the VM group (median [range], 2 [1–5] vs 3 [1–7], P < .001). The reason for stopping misoprostol was cited as sufficient cervical ripening more frequently for women in the VM group (39.5% vs 29.1%, P = .006). The maximum dose of oxytocin used during labor for women in the BM group was higher than for the VM group (6 mU/min vs 4 mU/min, P = .001). The overall rates of cesarean delivery were similar in the 2 groups (VM 15.8%, BM 22.3%, P = .15). There were no other statistically significant differences in treatment groups on delivery characteristics (Table 2) or in maternal or fetal serious adverse events (Table S1).
Participant preferences
There were no differences in participant satisfaction responses obtained the first or second day after delivery regarding their experience with induction of labor, expectations on pain, comfort regarding route of misoprostol, route they liked better, or preference of route in future IOL (Table S5). When asked which way of giving the medication was more comfortable, 43% responded they were the same, 39% said in the cheek, and 13% said in the vagina. When asked which dosing location they would prefer for a future induction if both medication routes were equivalent, 41.7% (36.1% to 47.3%) preferred “in my cheek,” 31.3% (26.1% to 36.6%) preferred “in my vagina,” and 21.7% (17.0% to 26.3%) said they were not sure.
Comment
Principal findings
In the IMPROVE trial, women receiving VM, compared with BM, were more likely to deliver more rapidly, deliver vaginally within 24 hours, and require fewer doses of misoprostol to achieve active labor. The BM group had significantly more cesarean deliveries for fetal nonreassurance. There were no differences in other adverse safety events.
Results in context
Several systematic reviews have compared alternative routes of misoprostol use to the vaginal route.13–15 Most have found similar effectiveness of nonvaginal routes. Buccal dosing of misoprostol cannot be assumed to be the same as the more studied sublingual dosing. Analyses of the pharmacokinetics of misoprostol given by buccal and sublingual routes clearly demonstrate differences.16 The onset of action of oral and sublingual routes are similarly fast (8–11 minutes) compared with vaginal route (20 minutes). However, most of the pharmacokinetic studies on misoprostol use high doses (600 mcg or more) not typically used for labor induction. The Cochrane Review comparing buccal to vaginal dosing of misoprostol identified only 1 trial of BM vs VM. However, it utilized different doses of misoprostol for each route of administration, making direct comparison impossible.14 The authors concluded that larger efficacy and safety trials were required to evaluate the buccal route. This has been accomplished in our trial.
Our rates of serious adverse events were low and were mostly due to prolonged hospitalization. Rates of tachysystole requiring therapeutic intervention were 14% in the vaginal group and 12% in the buccal group, similar to rates found in some other trials.12,17 This is in contrast to the trial by Carlan et al,10 which demonstrated higher rates of misoprostol-induced hyper-stimulation needing treatment in the buccal group (26%) than in the vaginal group (18%). However, that study used up to 300 mcg of buccal misoprostol but only 50 mcg vaginally.10
Clinical implications
As labor inductions are common, finding optimal methods to accomplish vaginal delivery is important. We have demonstrated that at typically used doses, VM leads to more rapid deliveries and has fewer urgent cesarean deliveries for fetal distress than buccal dosing. Given the recent findings of the ARRIVE trial,18 it is possible that the number of labor inductions will increase, further adding to the importance of optimizing this procedure.
Our subgroup analysis found that for multiparous women, vaginal dosing was superior to buccal dosing. This finding was true even after adjusting for other covariates such as the Bishop score at trial entry. This may be owing to biochemical differences in the cervices of nulliparous and multiparous women. Further investigation is warranted, as subgroup analyses are exploratory, but it is possible that residual uterine and cervical factors from prior deliveries may make multiparous women more responsive to VM. As nulliparous women frequently need cervical ripening, we plan to further explore the potential equivalence of the dosing routes in this group. A trial powered for nulliparous women is warranted.
Although we hypothesized that women undergoing labor induction would strongly prefer the medication by a nonvaginal route, we found similar rates of satisfaction with labor induction and dosing regimen preferences between groups. This is in contrast to previous, but nonblinded, studies that have found participants prefer sublingual dosing.12,19 Administering the study drug both buccally and vaginally in a blinded fashion, we found that preferences were not different. Interestingly, the rate of women preferring to have the medication in the cheek for a future induction was not much higher than those preferring the vagina (42% vs 31%, respectively). However, as the women were already anticipating cervical examinations, perhaps that led to a large number preferring vaginal dosing. In practice, however, when practitioners are not blinded to study drug as in our study, they may choose to skip some cervix examinations during cervical ripening with buccal dosing. This may have also influenced the route choice responses from participants. We did not specifically ask about the taste of the tablets, which also may have played a role in the responses.
Research implications
We are currently exploring potential pharmacokinetic differences in the 2 dosing routes and the role they may play in outcomes from misoprostol. Further exploration of the differences in response of nulliparous and multiparous women is also warranted as providers attempt to individualize labor induction methods. A systematic comparison of available dosing and routes of misoprostol for labor induction is warranted that would include buccal dosing.
Strengths and limitations
The major strength of our trial is that it had blinding of participants, providers, and study personnel/outcomes assessors (triple blinding), as well as blinding during data analysis, which reduces bias as compared to prior trials. To our knowledge, this is one of only a few trials utilizing BM in the same dose as VM for cervical ripening and labor induction at term for women with a live fetus. As many of our providers are uncomfortable placing vaginal drug in women with ruptured membranes, we enrolled only 1 woman with ruptured membranes. Thus, our findings are essentially limited to women with intact membranes. Our study was stratified by site, and differences between the populations at each site were accounted for in the analysis. But there could have been other differences in the 2 populations unaccounted for. Our findings of potential equivalence of the routes in nulliparous women is limited by this being a subgroup analysis not in the original power calculation.
Conclusions
In conclusion, we were unable to confirm noninferiority of BM vs VM. In fact, we found that VM may be superior to BM for cervical ripening at term. However, a randomized controlled trial specifically powered to detect a clinically meaningful difference from which to conclude superiority of VM to BM is still required. Vaginal dosing appears to lead to more rapid delivery and fewer cesarean deliveries for fetal distress.
Supplementary Material
AJOG at a Glance.
Why was this study conducted?
Buccal misoprostol is being used more commonly for cervical ripening but has not been well studied. Retrospective data indicated buccal misoprostol might be noninferior to vaginal misoprostol.
Key findings
We did not find buccal misoprostol to be noninferior to vaginal misoprostol. Women in the vaginal misoprostol group had higher rates of vaginal delivery, more rapid times from induction to delivery, and fewer cesarean deliveries for fetal heart rate abnormalities.
What does this add to what is known?
Vaginal misoprostol may be superior to buccal misoprostol. There were no differences in preference for buccal or vaginal misoprostol routes for women in the trial.
Acknowledgments
Content is solely the responsibility of the authors. The supporting organizations had no role in data acquisition, analysis or interpretation, manuscript creation, or the decision to submit for publication.
Footnotes
The authors declare no conflict of interest with the content of this manuscript. There was no funding source for this work other than internal Department funds for resident research. Dr Pierson’s time as an OB-Clinical Pharmacology was supported by NIH-NIGMS: Indiana University Comprehensive Training in Clinical Pharmacology (T32GM008425).
This trial is registered at clinicaltrials.gov (https://clinicaltrials.gov/ct2/show/NCT02408315) (ClinicalTrials.gov identifier: NCT02408315; first registered April 3, 2015; first participant enrolled October 4, 2015).
These results were presented at the 2018 Central Association of Obstetricians and Gynecologists Annual Meeting in Minneapolis, Minnesota, as the Central Prize Award Winning Abstract, October 2018.
References
- 1.Martin JA, Hamilton BE, Osterman M, Driscoll AK, Drake P. Births: final data for 2016. Nat Vital Stat Rep 2018;67:1–55. [PubMed] [Google Scholar]
- 2.ACOG Practice Bulletin No. 107: induction of labor. Obstet Gynecol 2009;114:386–97. [DOI] [PubMed] [Google Scholar]
- 3.Thomas J, Fairclough A, Kavanagh J, Kelly AJ. Vaginal prostaglandin (PGE2 and PGF2a) for induction of labour at term. Cochrane Database Syst Rev 2014:CD003101. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 4.Hofmeyr GJ, Gulmezoglu AM, Pileggi C. Vaginal misoprostol for cervical ripening and induction of labour. Cochrane Database Syst Rev 2010:CD000941. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 5.Krause E, Malorgio S, Kuhn A, Schmid C, Baumann M, Surbek D. Off-label use of misoprostol for labor induction: a nation-wide survey in Switzerland. Eur J Obstet Gynecol Reprod Biol 2011;159:324–8. [DOI] [PubMed] [Google Scholar]
- 6.Elati A, Weeks AD. The use of misoprostol in obstetrics and gynaecology. BJOG 2009;116: 61–9. [DOI] [PubMed] [Google Scholar]
- 7.Stephenson ML, Wing DA. Misoprostol for induction of labor. Semin Perinatol 2015;39: 459–62. [DOI] [PubMed] [Google Scholar]
- 8.Alfirevic Z, Keeney E, Dowswell T, et al. Methods to induce labour: a systematic review, network meta-analysis and cost-effectiveness analysis. BJOG 2016;123:1462–70. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9.Towns R, Quinney SK, Pierson RC, Haas DM. Survey of provider preferences regarding the route of misoprostol for induction of labor at term. AJP Rep 2017;7:e158–62. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 10.Carlan SJ, Blust D, O’Brien WF. Buccal versus intravaginal misoprostol administration for cervical ripening. Am J Obstet Gynecol 2002;186:229–33. [DOI] [PubMed] [Google Scholar]
- 11.Dorr ML, Pierson RC, Daggy J, Quinney SK, Haas DM. Buccal versus vaginal misoprostol for term induction of labor: a retrospective effectiveness cohort study [26K]. Obstet Gynecol 2016;127:96S. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 12.Nassar AH, Awwad J, Khalil AM, Abu-Musa A, Mehio G, Usta IM. A randomised comparison of patient satisfaction with vaginal and sublingual misoprostol for induction of labour at term. BJOG 2007;114: 1215–21. [DOI] [PubMed] [Google Scholar]
- 13.Alfirevic Z, Aflaifel N, Weeks A. Oral misoprostol for induction of labour. Cochrane Database Syst Rev 2014;6:CD001338. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14.Muzonzini G, Hofmeyr GJ. Buccal or sublingual misoprostol for cervical ripening and induction of labour. Cochrane Database Syst Rev 2004:CD004221. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 15.Souza AS, Amorim MM, Feitosa FE. Comparison of sublingual versus vaginal misoprostol for the induction of labour: a systematic review. BJOG 2008;115: 1340–9. [DOI] [PubMed] [Google Scholar]
- 16.Schaff EA, DiCenzo R, Fielding SL. Comparison of misoprostol plasma concentrations following buccal and sublingual administration. Contraception 2005;71:22–5. [DOI] [PubMed] [Google Scholar]
- 17.Caliskan E, Bodur H, Ozeren S, Corakci A, Ozkan S, Yucesoy I. Misoprostol 50 microg sublingually versus vaginally for labor induction at term: a randomized study. Gynecol Obstet Invest 2005;59:155–61. [DOI] [PubMed] [Google Scholar]
- 18.Grobman WA, Rice MM, Reddy UM, et al. Labor induction versus expectant management in low-risk nulliparous women. N Engl J Med 2018;379:513–23. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 19.Zahran KM, Shahin AY, Abdellah MS, Elsayh KI. Sublingual versus vaginal misoprostol for induction of labor at term: a randomized prospective placebo-controlled study. J Obstet Gynaecol Res 2009;35: 1054–60. [DOI] [PubMed] [Google Scholar]
Associated Data
This section collects any data citations, data availability statements, or supplementary materials included in this article.