Objective
The purpose of this study was to review randomized trials that compared the efficacy of the dinoprostone 10-mg controlled-release vaginal insert with that of vaginally administered misoprostol tablets in cervical ripening and labor induction.
Study Design
Electronic databases were used to identify randomized clinical trials that compared dinoprostone with misoprostol. Estimates of the relative risk for individual studies and risk difference from dichotomous outcomes with the use of random and fixed-effects models were calculated.
Results
Women who received misoprostol had a higher incidence of vaginal delivery within 12 and 24 hours of prostaglandin application, compared with dinoprostone. Both modalities had similar incidences of cesarean delivery, uterine hyperstimulation, and fetal tachysystole. There was an increased need for oxytocin augmentation in the dinoprostone group. No significant difference in neonatal outcomes was noted between the 2 groups.
Conclusion
Vaginally administered misoprostol was more effective than the dinoprostone vaginal insert for cervical ripening and labor induction. The safety profiles of both drugs were similar.
Induction of labor is an increasingly common practice that accounts for >20% of all births. Methods for labor induction include both mechanical and pharmacologic options. Although oxytocin is an effective drug for the augmentation of labor in patients with favorable cervices, in the patient with an unfavorable cervix, a ripening agent may be used. In 1995, the Food and Drug Administration approved a dinoprostone vaginal insert (Cervidil; Forest Pharmaceuticals Inc, St. Louis, MO) for cervical ripening. This insert releases 10 mg of dinoprostone over 12 hours and was shown to shorten the interval between induction and vaginal delivery safely. Before the Food and Drug Administration approval, most data that assessed safety and efficacy of the dinoprostone vaginal insert were based on placebo-controlled trials. Although comparisons with placebo showed encouraging results, the dinoprostone vaginal insert was not compared with other available prostaglandin preparations. Recent clinical trials have compared the safety and efficacy of dinoprostone vaginal inserts with other active agents, mainly intravaginal and intracervical dinoprostone gel, intravaginal misoprostol tablets, and misoprostol vaginal inserts. As research regarding the optimal method of cervical ripening continues, the growing body of literature warrants reassessment. This metaanalysis of randomized controlled trials compares the efficacy and safety of the dinoprostone 10-mg vaginal insert as a cervical-ripening and labor-induction agent with the intermittent vaginal administration of misoprostol tablets.
Materials and Methods
Because this was a metaanalysis, it was exempt from institutional review board approval. We used several sources to identify all published and unpublished randomized clinical trials that compared dinoprostone vaginal inserts for cervical ripening and labor induction with vaginal misoprostol. We conducted searches from 1990 through March 2009 using Bibliographic Retrieval Service (BRS Online, MacLean, VA), MEDLINE (National Library of Medicine, Bethesda, MD), Current Contents (Institute for Scientific Information, Philadelphia, PA), and Silver Platter (Silver Platter Information Inc, Norwood, MA) databases. Medical subject heading search terms included labor induction, cervical ripening, prostaglandins, misoprostol , and dinoprostone vaginal insert .
We manually screened references of studies and review articles for relevant citations without regard to language of publication. Abstracts and review articles were analyzed. Only randomized controlled trials that compared 10-mg dinoprostone vaginal inserts with misoprostol for cervical ripening and labor induction were included in the metaanalysis. Studies were included without regard to status of membranes or gestational age. Each study’s methodologic attributes were assessed for quality by 2 investigators, and a mean quality score was tabulated. The materials and methods and results sections of each study were photocopied separately after identifiers were removed, so the investigators were masked to study authors and institutions. Each study was graded 0-9 for randomization, blinding, explicit inclusion criteria, explicit exclusion criteria, baseline similarity between groups for important prognostic features, equivalent and comparable administration of vaginal inserts within and among trials, equal detection of outcomes between groups, and appropriate statistical analysis. Mean quality scores were not used for including studies.
The main outcomes, which were selected before we reviewed individual studies, included the proportion of women who delivered vaginally within 12 and 24 hours of drug application, the need for oxytocin augmentation, the incidence of hyperstimulation (at least 6 uterine contractions per 10-minute period with fetal heart rate [FHR] abnormalities), tachysystole (at least 6 uterine contractions per 10-minute period without fetal heart rate abnormalities), and the proportion of women who delivered by cesarean section. We also assessed the impact of these labor induction agents on the rates of neonatal intensive care unit (NICU) admissions and Apgar scores <7 at 5 minutes after birth. The effect of vaginal inserts on cervical scores was not reported in most studies. Not all trials in the metaanalysis evaluated each of those outcomes; therefore, specific outcome analyses were based on a variable number of studies.
Data from trial reviews were analyzed with Stata statistical software package (version 8; Stata Corporation, College Station, TX) and included descriptive statistics and relative risk (RR,) with 95% confidence intervals (CIs) for individual studies. We calculated an estimate of the RR, and risk difference from dichotomous outcomes using a random- (DerSirmonian and Laird) and fixed-effects (Mantel and Haenszel) model. Random effects models were used whenever there was evidence of significant heterogeneity. The null hypothesis underlying the overall test of association was that the overall RR, was equal to 1. To determine the combinability of individual studies, we did a formal test of homogeneity of treatment effect across the studies. In a further examination of homogeneity of effects, we plotted data with the rate of outcome variables in the control groups on the X axis and in the dinoprostone vaginal insert group on the Y axis, as suggested by L’Abbe et al.
Publication bias was examined through visual inspection of a funnel plot, whereby RRs were plotted against study sample size. Sensitivity analyses were performed by the sequential omission of each study and an analysis the overall impact on the pooled results.
Results
We identified 27 studies that assessed effectiveness of 10-mg dinoprostone vaginal inserts for cervical ripening and labor induction, 16 of which were excluded ( Figure ) : 1 study because it was a noncomparative surveillance study that reported the experience at a community hospital; 2 studies because they were basic science research studies that assessed pharmacokinetics of the vaginal inserts; 6 studies because they compared dinoprostone with prostaglandins other than misoprostol; 3 studies because the type of pessary used was distinct from the controlled-release vaginal insert; 1 study because a misoprostol vaginal insert was used rather than an intravaginal tablet, and the last 3 studies because the vaginal insert was compared with placebo.
Eleven randomized clinical trials that compared the 10-mg, sustained-release dinoprostone vaginal insert with vaginal misoprostol tablets for cervical ripening and labor induction were evaluated further. Mean quality scores for those studies were similar. A total of 1572 women were enrolled in those trials (785 in the vaginal insert group and 787 in the misoprostol group). The number of women in the various vaginal insert groups ranged from 38-115 women, with comparison groups generally of similar size (38-108 women). Of 787 women who were allocated to misoprostol groups, 550 women (69.9%) received 50 μg, and 237 (30.1%) received ≤35 μg. Design features of the selected trials are shown in Table 1 . Despite differences in inclusion criteria, most of the women had intact membranes and were >36 weeks of gestation at the time of induction. Women also had no major differences in maternal age, parity, gestational age, history of cesarean deliveries, or indication for induction. Quality assessment indicated that each trial was adequate for inclusion in the metaanalysis. Three trials compared vaginal inserts with intravaginal misoprostol at a dose of 25 μg. Eight trials compared the vaginal inserts with intravaginal misoprostol at a dose of 50 μg. Eight studies were performed in the United States; 2 studies were from Turkey, and 1 study was from France. The studies differed in size, inclusion criteria, and dosing protocols of both the dinoprostone vaginal insert and the intravaginal misoprostol tablets. Of particular importance was the range of Bishop scores for inclusion among the various included trials. Although different Bishop scores were used to determine inclusion criteria, 10 of the 11 studies had Bishop scores of ≤6 in both the dinoprostone and the misoprostol groups. Most studies had inclusion criteria Bishop scores between 4 and 6; 1 study did not have a stated Bishop score requirement for enrollment. However, initial Bishop scores were still documented in this study, and >85% of participants had a Bishop score of ≤6. All but 2 studies required intact membranes for enrollment.
| Author (y, country) | Participants | Dinoprostone dose | Misoprostol dose | Percentage of nulliparity | Quality score | Inclusion criteria | Methods |
|---|---|---|---|---|---|---|---|
| Wing et al (1997, USA) | 197 total: 98 dinoprostone; 99 misoprostol | 1-mg dinoprostone vaginal insert up to 24 h | 25 μg per vagina every 4 h to 150 μg | Dinoprostone, 42.9%; misoprostol, 48.5% | 3 | S, C, I, B ≤4, N, H ≤8/1 h | Agent removed/stopped for spontaneous rupture of membranes, active labor, adequate ripening, or fetal heart rate/ contractile abnormality; oxytocin used 2 h after removal/last dose as needed for inadequate contractions |
| Sanchez-Ramos et al (1998, USA) | 223 total: 115 dinoprostone; 108 misoprostol | 10-mg dinoprostone vaginal insert up to 12 h | 50 μg per vagina every 3 h to 24 h or 400 μg | Dinoprostone, 56.5%; misoprostol, 60.2% | 2 | O, M, A, CD, V, S, C | Agent removed/stopped for adequate labor; insert removed for tachysystole/ hyperstimulation; oxytocin started for inadequate contractions; artificial rupture of membranes when appropriate; internal monitors placed |
| Harms et al (2001, USA) | 121 total: 60 dinoprostone; 61 misoprostol | 10-mg dinoprostone vaginal insert to 12 h | 25 μg per vagina, then N/A 25-50 μg every 4 h to 3 doses | Not available | I, B ≤6, T | Oxytocin started after both agents | |
| Bolnick et al (2004, USA) | 151 total: 74 dinoprostone; 77 misoprostol | 10-mg dinoprostone vaginal insert to 12 h | 25 μg per vagina every 4 h to dinoprostone, 54.1%, to 6 doses misoprostol, 52.0% | 3 | T, S, C, B ≤6, I, N, H <4/20 min | Agent removed/stopped for spontaneous rupture of membranes or regular contractions; misoprostol stopped after cervix >4 cm dilated; oxytocin given 2 h after last misoprostol dose as needed, given concurrently with insert at standard low dose | |
| Bebbington et al (2003, USA) | 200 total: 102 dinoprostone; 98 misoprostol | 10-mg dinoprostone vaginal insert to 12 h | 50 μg per vagina every 4 h to 3 doses | Dinoprostone, 62.7%; misoprostol, 53.1% | 2 | S, T, C, B ≤6 | Dinoprostone removed for uterine hypertonus/hyperstimulation, active labor, or after 12-h period; misoprostol dosing stopped for frequent contractions (3/10 min) |
| Khourry et al (2001, USA) | 118 total: 39 dinoprostone; 40 misoprostol a | 10-mg dinoprostone vaginal insert to 22.5 h | 35-μg and 50-μg vaginal suppositories every 4.5 h to 210 μg (35 μg group) or 300 μg (50 μg group) | Dinoprostone, 59.0%; misoprostol, a 50.0% | 3 | O, S, T, C, I, B ≤4, N | |
| Garry et al (2003, USA) | 186 total: 89 dinoprostone; 97 misoprostol | 10-mg dinoprostone vaginal insert every 12 h to 24 h | 50 μg every 3 h to 400 μg | Dinoprostone, 67%; misoprostol, 61% | 3 | O, B ≤4, C, N | Dinoprostone repeated if cervix not ripe after 12 h; misoprostol held for ≥3 contractions in 10 min |
| Rozenberg et al (2004, France) | 140 total: 70 dinoprostone; 70 misoprostol | 10-mg dinoprostone vaginal insert for 12 h; course repeated on second day | 50 μg per vagina every 6 h to 100 μg on day 1, then 50 μg every 4 h to 150 μg on second day as needed | Dinoprostone, 67.1%; misoprostol, 62.9% | 2 | M, B ≥5, S, C, H <4/h | |
| Tanir et al (2008, Turkey) | 87 total: 44 dinoprostone; 43 misoprostol | 10-mg dinoprostone vaginal insert every 12 h to 24 h | 50 μg per vagina every 6 h; not given to 4 doses | 3 | T, S, C, I, B <6, A | If Bishop score ≥5, oxytocin started followed by artificial rupture of membranes | |
| Ozkan et al (2009, Turkey) | 112 total: 56 dinoprostone; 56 misoprostol | 10-mg dinoprostone vaginal insert ×12 h | 50 μg per vagina every 4 h to 250 μg | Dinoprostone, 57.1%; misoprostol, 51.8% | 2 | S, T, O, C, I, B ≤4, N | |
| Ramsey et al (2003, USA) | 111 total: 38 dinoprostone; 38 misoprostol | 10-mg dinoprostone vaginal insert × 12 h | 50 μg per vagina every 6 h × 2 doses | Dinoprostone, 57.9%; misoprostol, 60.5% | 4 | B ≤5, S, C, V, H = 4/h, N | |
| Wing a (2008, USA) | 879 total: 436 dinoprostone; 443 misoprostol | 10-mg dinoprostone vaginal insert | 50 μg vaginal insert | Dinoprostone, 61.7%; misoprostol, 62.4% | 4 | 18+, LP, S, T, B ≤4, no UD | Misoprostol was stopped when patient reached active labor, had maternal or fetal complication, reached 24-h dosage, or spontaneously fell out |
Among women who were assigned randomly to dinoprostone, the likelihood of vaginal delivery within 12 hours was lower (RR, 0.65; 95% CI, 0.44–0.96; Table 2 ). The likelihood of vaginal delivery within 24 hours was also lower (RR, 0.83; 95% CI, 0.74–0.94; Table 3 ). More participants who underwent induction with dinoprostone required oxytocin augmentation (RR, 1.45; 95% CI, 1.20–1.74; Table 4 ). No statistical differences between the 2 groups were noted with respect to rates of cesarean delivery (RR, 1.01; 95% CI, 0.85–1.19; Table 5 ), rates of fetal tachysystole (RR, 0.91; 95% CI, 0.53–1.54; Table 6 ), or rates of hyperstimulation (RR, 0.86; 95% CI, 0.59–1.25; Table 7 ). Although use of the dinoprostone insert was associated with an increased likelihood of cesarean delivery (RR, 1.01), this difference did not achieve statistical significance. In addition, the likelihood of 5-minute Apgar scores <7 (RR, 0.99; 95% CI, 0.41–2.36; Table 8 ) or NICU admissions (RR, 0.93; 95% CI, 0.68–1.29; Table 9 ) were similar in the 2 groups.
