Conflict of interest: All of the authors have disclosed no financial relationship with a biotechnology manufacturer, a pharmaceutical company or other commercial entity that has an interest in the subject matter or materials discussed in the manuscript.
Recent studies suggest that misoprostol may be more effective than dinoprostone in pregnant women with unfavorable cervix. The objective here is to investigate and compare the efficacy and safety of intravaginal misoprostol and intracervical dinoprostone for labor induction, including incidence of cesarean section, vaginal delivery rate within 24 h, uterine hyperstimulation, tachysystole, oxytocin augmentation, neonatal intensive care unit (NICU) admissions, and Apgar score of less than 7 at 1 and 5 min.
Databases searched were MEDLINE, EMBASE and Cochrane Central Register of Controlled Trials, up to July 2013. Randomized controlled trials comparing intravaginal misoprostol with intracervical dinoprostone in women with singleton pregnancy, intact membranes and unfavorable cervix (Bishop's <6) were included. Pooled relative risk, mean difference and 95% confidence intervals were calculated.
The use of misoprostol was significantly effective in increasing the rate of vaginal delivery within 24 h and less oxytocin augmentation when compared with dinoprostone. However, the incidents of uterine hyperstimulation and tachysystole were significantly higher under the misoprostol protocol than dinoprostone protocol. Furthermore, we found similar efficiency in the rate of cesarean delivery, NICU admission and Apgar score at 1 and 5 min among the study groups.
Intravaginal misoprostol appears to be more efficient for labor induction than intracervical dinoprostone; however, dinoprostone has been demonstrated to be safer because of the lower incidence of uterine hyperstimulation and tachysystole. Further high-quality studies assessing the possible effectiveness of misoprostol and dinoprostone in selected groups of patients are warranted.
Induction of labor (IOL) is an increasingly common obstetric procedure. Recent data demonstrates that over 40% of primiparous women, and over 30% of multiparous women, undergo labor induction. Methods for labor induction include both mechanical and pharmacological options. Pharmacological interventions to ripen the cervix as part of labor induction include administration of oxytocin, and prostaglandins delivered p.o. or vaginally. However, when IOL is attempted for a woman with an unfavorable cervix, other interventions used to assist the induction process, such as oxytocin or rupture of membranes, are connected with reduced effectiveness and high failure rates.
Misoprostol is a prostaglandin E1 analog that was first marketed in the 1980s to prevent and treat peptic ulcer disease. However, due to its effects on uterine contractility and cervical ripening, a number of randomized trials and systematic reviews have assessed its use in obstetric and gynecologic procedures. Misoprostol's low cost, stability in a wide range of temperature and availability in over 80 countries make it particularly useful in resource-poor settings. Labor induction with dinoprostone has otherwise been chiefly used with either a 0.5-mg dinoprostone gel delivered intracervically, known as Prepidil (released with US Food and Drug Administration [FDA] approval in 1992 by Pfizer) or a 10-mg controlled-release vaginal insert known as Cervidil (released with FDA approval in 1995 by Forest Laboratories). Dinoprostone is expensive and requires cold storage to keep the compound chemically stable, which impacts ease of use.
Hofmeyr et al. had reported a Cochrane review comparing vaginal administration of misoprostol with vaginal or intracervical dinoprostone on cervical ripening or labor induction. Placebo and oxytocin were used in the study. The meta-analysis included 121 randomized controlled trials (RCT), of which 13 were double-blinded. Vaginal administration of misoprostol compared with vaginal or intracervical administration of dinoprostone gel or i.v. oxytocin was associated with a higher rate of vaginal delivery within 24 h and less oxytocin augmentation, but more uterine hyperstimulation with and without FHR changes. However, in this review, multiple pregnancies, previous cesarean section, premature rupture of membranes and favorable cervix were all included, which would have influenced the results extensively. Meanwhile, after this research, more studies reported results about misoprostol and dinoprostone. The growing body of published work warrants reassessment.
To explore this clinically crucial domain, our objective was to conduct a methodologically valid systematic review and meta-analysis of RCT comparing intravaginal misoprostol with intracervical dinoprostone in singleton pregnant women with unfavorable cervix at term.
Published work search
An expert reference librarian (Y. Hu) projected and conducted the electronic search strategy in consultation with an expertise of obstetrics in performing systematic reviews. The following electronic databases were searched: MEDLINE (1950 to July 2013), EMBASE (1980 to July 2013), the Cochrane Central Register of Controlled Trials (CENTRAL) (issue 7, 2013). Search terms were applied as follows: (‘Misoprostol’ OR ‘SC-30249’ OR ‘Cytotec’ OR ‘Prostaglandins’ OR ‘Misoprostol Apotex Brand’) AND (‘Dinoprostone’ OR ‘PGE2 alpha’ OR ‘PGE2’ OR ‘Prostaglandin E2’ OR ‘Prepidil Gel’ OR ‘Prostaglandins E’ OR ‘Prostaglandins’ OR ‘Prostenon’) AND (‘cervical ripenings’ OR ‘uterine cervix ripening’ OR ‘induced labor’ OR ‘labor induction’). The reference lists of included articles were also searched manually. When necessary, additional information was sought from the authors. The search was restricted to the English language.
Inclusion and exclusion criteria
To be included in the meta-analysis, published RCT comparing intravaginal misoprostol with intracervical dinoprostone had to meet the following inclusion criteria: (i) nulliparous and/or multiparous women admitted for the induction of labor at term (>37 weeks); (ii) singleton pregnancy with cephalic presentation and no contraindication to vaginal delivery; (iii) unfavorable cervix (Bishop's <6); (iv) intact membranes; (v) absence of active labor or fetal distress; (vi) at least one of the following outcomes was assessed: incidence of cesarean section, vaginal delivery rate within 24 h, uterine hyperstimulation, tachysystole (TS), oxytocin augmentation and neonatal intensive care unit (NICU) admissions, and Apgar score less than 7 at 1 and 5 min.
Exclusion criteria included: (i) ruptured membranes; (ii) previous cesarean delivery or history of uterine surgery; and (iii) cephalopelvic disproportion. In addition, trials without control groups, animal trials, reviews and non-randomized trials were excluded from consideration. The study selection was performed independently by two reviewers (J. Lang and J. Lv). Any disagreements were settled by consultation with a third reviewer (A. Liu).
Quality assessment and data extraction
The methodological quality of the included trials was assessed using the Cochrane Handbook for Systematic Reviews of Interventions. An aggregate score between 0 (weakest) and 8 (strongest) was evaluated for each included trial.
Data on study design (including the intervention and comparator) and characteristics of study patients (including number of subjects enrolled, mean or median age, gestational age and initial Bishop score) were extracted from qualifying studies. Disagreements were resolved through consensus or discussed with authority (A. Liu).
Data entry and statistical analysis were performed by using Review Manager software (RevMan 5.2; Cochrane Collaboration, Oxford, UK) by fixed-effects method. When substantial heterogeneity existed, the random-effects method was used.
The relative risk (RR) combined with 95% confidence intervals (95% CI) were used to present the effect estimate for dichotomous data. On the other hand, in case of continuous data, we used the mean difference (MD) to assess. Statistical significance was set at P < 0.05.
Statistical heterogeneity was identified by I2 (>50%) and χ2-test (P < 0.1), which estimated the appropriateness of pooling the individual study results.
The study selection process is detailed in Figure 1. A total of 1651 studies were retrieved in the initial electronic search. After removal of duplicates, 1290 records remained. After title and abstract screening, 1224 studies were excluded. The remaining 66 studies were considered adequate by one or both reviewers. During the second phase of the inclusion process, 56 were excluded for the following reasons: (i) 42 articles compared intravaginal/intracervical/oral/sublingual misoprostol with intravaginal dinoprostone; (ii) three articles included premature rupture of membranes; (iii) six articles included gestational age of less than 37 weeks or Bishop's score of more than 6; and (iv) five articles were reviews.
Description of included studies
Ten RCT[8-17] evaluating intravaginal misoprostol versus intracervical dinoprostone for labor induction in pregnant women at term met the inclusion criteria. The sample size ranged 44–200 women with a total of 1061 women. Of these 10 trials, six articles administrated 50 mcg misoprostol intravaginally, while four studies administrated 25 mcg. Meanwhile, all included articles used 0.5 mg dinoprostone gel intracervically. Nine studies received oxytocin for augmentation when there were no changes in cervical dilatation for more than 2 h in the active phase of labor. Characteristics of the articles included in the meta-analysis are shown in Table 1.
Table 1. Characteristics of included studies
Study, year (reference)
Mean age (year ± SD)
Gestational age (weeks)
Initial bishop score
Misoprostol group treatment
Dinoprostone group treatment
D, dinoprostone; M, misoprostol; SD, standard deviation.
25 mcg intravaginal repeated every 6 h, maximum 50 mcg
0.5 mg intracervical repeated every 6 h, maximum 1.0 mg
Risk of bias in included studies
As summarized in Table 2, we assessed the risk of bias with the Cochrane Handbook for Systematic Reviews of Interventions. The number of criteria corresponding ranged from two to five of the eight criteria. Seven included studies described the method of random sequence generation.[8-11, 13, 14, 16] Five trials reported allocation concealment.[8-11, 13] No double-blinding was found in any of the 10 RCT, five studies reported the blinding of participants, one article described blinding of the investigator and one study described blinding of the assessor. One article reported withdrawal and intention-to-treat analyses, the remainder had no dropout. Non-selective reporting presented in three trials.[8, 10, 13] Five studies had a certain degree of other potential threats to validity. Except for the clinical studies of Ramsey et al. and Nanda et al., the risk of bias was high or medium for the majority of trials.
Table 2. Risk of bias of included studies
Studies fulfilling the criteria: A, adequate sequence generation; B, concealment of allocation; C, blinding (patient); D, blinding (investigator); E, blinding (assessor); F, incomplete outcome data addressed (ITT analysis); G, free of selective reporting; H, other potential thereat to validity. +, yes; –, no; ?, unclear.
Three studies[14, 16, 17] with a total of 500 women reported data for the outcome of clinical pregnancy. Vaginal delivery at less than 24 h occurred in 153 of 235 (65.1%) women randomized to misoprostol and in 135 of 265 (50.9%) women randomized to dinoprostone. The use of misoprostol was significantly effective in increasing the rate of vaginal delivery at less than 24 h compared with dinoprostone (RR = 1.27; 95% CI, 1.10–1.48; P = 0.002; Fig. 2). No heterogeneity existed (P = 0.70, I2 = 0%).
The effect of misoprostol and dinoprostone on cesarean delivery was evaluated in 10 studies for a total of 1061 patients. Cesarean delivery emerged in 132 of 537 (24.6%) women randomized to misoprostol and in 135 of 524 (25.8%) randomized to dinoprostone. The use of misoprostol did not result in a statistically significant higher rate of cesarean delivery as compared with dinoprostone (RR = 0.95; 95% CI, 0.78–1.17; P = 0.66; Fig. 3). No significant statistical heterogeneity was detected (P = 0.23, I2 = 23%).
Nine studies reported data on development of oxytocin augmentation. With a total of 988 patients, 172 out of 499 (34.5%) women randomized to misoprostol, and 270 out of 489 (55.2%) randomized to dinoprostone, needed oxytocin augmentation. This demonstrated that there was less statistical significance for administration of oxytocin for misoprostol compared with dinoprostone (RR = 0.62; 95% CI, 0.54–0.72; P < 0.00001; Fig. 4). No significant heterogeneity was suggested (P = 0.10, I2 = 40%).
The rate of uterine hyperstimulation was higher for women treated with misoprostol (31/401, 7.7%) than for those treated with dinoprostone (9/391, 2.3%), in which the pooled analysis of eight trials was reported (RR = 3.15; 95% CI, 1.58–6.28; P = 0.001; Fig. 5). No significant heterogeneity existed in this comparison (P = 0.54, I2 = 0%).
Pooled analysis of six studies that estimated the incidence of TS, in association with misoprostol and dinoprostone in each study group, showed that there was a significantly higher incidence of TS in misoprostol protocol, when compared with dinoprostone (RR = 2.02; 95% CI, 1.28–3.19; P = 0.003; Fig. 6), and there was no heterogeneity among studies (P = 0.11, I2 = 44%).
Five studies provided data on the incidence of NICU admission. No significant differences between the misoprostol group and dinoprostone group were observed (RR = 0.95; 95% CI, 0.62–1.45; P = 0.80; Fig. 7). No heterogeneity emerged in this comparison (P = 0.59, I2 = 0%).
Apgar score at 1 and 5 min
The mean (±standard deviation [SD]) data of Apgar score at 1 and 5 min were both similar for pregnant women treated with misoprostol compared with those treated with dinoprostone, in which the pooled analysis of three trials was administered (1 min: MD, 0.03; 95% CI, −0.38 to 0.43; P = 0.90; 5 min: MD, 0.02; 95% CI, −0.33 to 0.38; P = 0.89; Fig. 8). Significant heterogeneity existed at 5 min, so the random-effects model was used (P = 0.13, I2 = 56%; Fig. 8), and no heterogeneity emerged at 1 min (P = 0.23, I2 = 31%; Fig. 8).
Meanwhile, we also reported four studies on the rate of Apgar score of less than 7 at 1 min and 5 min, and no statistically significant difference occurred between the two groups (1 min: RR = 0.99; 95% CI, 0.61–1.60; P = 0.96; 5 min: RR = 0.58; 95% CI, 0.19–1.76; P = 0.34; Fig. 8). No heterogeneity was found (1 min: P = 0.37, I2 = 5%; 5 min: P = 0.56, I2 = 0%; Fig. 8).
This study is the first meta-analysis regarding the therapeutic effects of intravaginal misoprostol and intracervical dinoprostone in singleton pregnant women with unfavorable cervix at term. Intravaginal misoprostol has higher efficacy than intracervical dinoprostone with regards to vaginal delivery at less than 24 h and oxytocin augmentation while the incidences of uterine hyperstimulation and TS were shown to be significantly higher in the misoprostol than dinoprostone groups. Furthermore, we found no differences in the rate of cesarean delivery, NICU admission, and Apgar score at 1 and 5 min among the study groups.
This study had some limitations. The methodological quality of the included 10 RCT was generally weak, and only two studies[10, 13] presented a low risk of bias. Most included articles did not report whether or not allocation concealment was fully attained and blinding requirements were fully met. Hence, selective bias may have existed in the trials. Paucity of double blinding(s) is prone to potential bias due to physician and patient awareness of treatment allocation. Studies have shown that trials with inadequate allocation concealment overstated treatment efficacy by 30–40% compared with trials that had adequate allocation concealment. Non-selective reporting was found only in three trials,[8, 10, 13] meanwhile, our inclusion criterion limited language to English, with studies that indicated a statistically significant effect of treatment being more likely to be published,[18, 29, 30] published in English, cited by other authors and generate multiple publications[33, 34] than other studies. Such studies are also more likely to be brought into systematic reviews, which may therefore overestimate the beneficial effects of treatment, ultimately leading to language bias and publication bias.
The following factors have strengthened this meta-analysis. Studies were included or excluded according to strict criteria. Women with ruptured membranes, previous cesarean delivery, favorable cervix and gestational age of less than 37 weeks were excluded without hesitation, and those who used intravaginal dinoprostone were also excluded. In addition, the heterogeneity of results used in the different studies makes the overall interpretation of data difficult. Fortunately, except for the mean (±SD) data of Apgar score at 5 min, which used the fixed-effects model, no heterogeneity existed in other outcomes.
Misoprostol increased vaginal delivery within 24 h and reduced need for oxytocin augmentation when compared with dinoprostone; meanwhile, we found that misoprostol was similar to dinoprostone with respect to cesarean delivery, which are coincident with other systematic reviews comparing misoprostol with other prostaglandins for cervical ripening and labor induction. In most countries, doctors are prone to use dinoprostone in the form of a vaginal insert for unprepared cervix. However, both vaginal insert and gel are recognized to be equally effective in overall time to mode of birth and neonatal outcome. Misoprostol has been used for induction of labor since the 1990s. The World Health Organization expert committee included 25 μg intravaginal misoprostol in the complementary list of the model list for the induction of labor at-term. The International Federation of Gynecology and Obstetrics also recently recommended intravaginal administration of 25 μg misoprostol every 4 h for a maximum of six doses for induction of labor at term. Hofmeyr et al. suggested that lower doses of misoprostol compared with higher doses did not reveal significant differences except for more need for oxytocin augmentation. Others[8, 22, 23] found that oxytocin augmentation was used more often in the 25-μg dose group than higher dose. Owing to various doses and various routes of misoprostol included in our review, subgroup analysis seemed to be unsuitable. A Cochrane review reported that a 25-mcg dose of misoprostol was associated with less uterine hyperstimulation when compared with a 50-mcg dose.
American Congress of Obstetricians and Gynecologists warns that higher doses (50 mcg every 6 h) of misoprostol could result in more uterine hyperstimulation and FHR decelerations, and recommends a 25-mcg dose of misoprostol inserted into the posterior vaginal fornix, repeated every 3–6 h as needed. It seems that a 25-mcg dose of misoprostol may be the optimal treatment; more and larger studies comparing different doses are necessary to demonstrate this viewpoint.
The incidence of uterine hyperstimulation and TS were higher in the misoprostol protocol than dinoprostone protocol; however, similar efficacy happened in NICU admission and Apgar score at 1 and 5 min. The preferred term, ‘tachysystole’ (TS), is defined as more than five contractions in 10 min, averaged over a 30-min window. A very recent study which focused on TS during early labor of women who received misoprostol, found that TS was associated with decelerations, but not adverse infant outcomes. Heuser et al. indicated that TS events increased the rate of NICU admission and composite adverse neonatal outcome by approximately 30%. Uterine hyperstimulation has been associated with abnormal FHR patterns due to inadequate uterine relaxation time. Physiologically, the frequent compression of the uterine spiral arterioles without adequate relaxation time is thought to lead to diminished placental perfusion and destroyed delivery of oxygen to the fetus, increasing the likelihood of fetal hypoxia and acidosis.
In conclusion, intravaginal misoprostol appears to be more efficient for labor induction than intracervical dinoprostone; however, dinoprostone seems to be more safe, due to the lower rate of uterine hyperstimulation and TS. In the future, larger and meticulously designed clinical trials will be indispensable for better definition.