, Yun Wang 1,*
1 Department of Obstetrics and Gynecology, Hangzhou Women’s Hospital, 310008 Hangzhou, Zhejiang, China
Abstract
Labor induction is a common procedure in obstetrics, with various methods employed to facilitate the process. This study aimed to compare the effects of labor induction using the double-balloon catheter (DBC) and dinoprostone (D), two widely used methods in clinical practice.
Randomized controlled trials (RCTs) were identified from the Cochrane Library, SinoMed, PUBMED, and Embase up to March 2024. Two researchers independently screened and extracted the literature according to the inclusion and exclusion criteria. After evaluating the literature, Revman 5.4 software was used to analyze the data. The time to the onset of active labor and the rate of cesarean deliveries were the key outcomes. The secondary outcomes included assisted vaginal delivery, failure to progress rate, failed induction rate, incidence of nonreassuring fetal heart rate, neonatal asphyxia incidence, neonatal intensive care unit (NICU) admission rate, improvement in time to onset of active labor, time to vaginal delivery, Bishop score, need for oxytocin administration, presence of postpartum hemorrhage, vaginal delivery within 24 hours, and uterine hyperstimulation rate.
The total of 7 RCTs involving 2267 patients were included, with 1127 treated with DBC, and 1140 treated with D. Our meta-analysis showed no significant difference between the two groups in terms of cesarean delivery [odds ratio (OR) = 1.10, 95% confidence interval (95% CI): 0.91, 1.32, p = 0.34], time to onset of active labor [mean difference (MD) = 0.66, 95% CI: –4.72, 6.03, p = 0.81], presence of nonreasuring fetal heart, success rate of induction, improvement in Bishop score, rate of progress in labor, vaginal delivery rate within 24 hours, time to vaginal delivery, postpartum hemorrhage, and assisted vaginal delivery. Although the oxytocin administration rate (OR = 3.96, 95% CI: 3.18, 4.93, p < 0.00001) was lower, D was more likely to cause uterine hyperstimulation, leading to a higher rate of neonatal asphyxia (OR = 0.39, 95% CI: 0.18, 0.85, p = 0.02).
Compared to DBC, D was more likely to cause uterine hyperstimulation, leading to a higher rate of neonatal asphyxia.
Keywords
- double-balloon catheter
- dinoprostone
- induction of labor
Induction of labor (IOL) is a common procedure for pregnant patients. It implies the use of artificial methods to promote cervical maturation and induce uterine contractions to achieve the goal of delivery. According to the National Vital Statistics report, in the United States, 31.9% of women underwent IOL in 2022 [1]. The plan of induction depends on the cervical Bishop score which was developed by Dr. Edward Bishop in 1964 [2]. It is comprised of fetal head station, cervical consistency, position, effacement and dilation. When the modified Bishop score is less than 7, preinduction cervical ripening—which involves utilizing pharmaceutical or mechanical methods to dilate, efface and soften the cervix—should be considered, with the exception of term premature rupture of fetal membranes (PROM) [3].
Cervical ripening is mainly divided into a mechanical or pharmaceutical option. Mechanical option for IOL uses physical effects to dilate the cervix. The mechanical option can stimulate the release of endogenous prostaglandins (PGs) from the cervix by increasing the separation of the amnion from the fetal-maternal interface [4]. Foley catheter and double-balloon catheter (DBC) are 2 of the most common mechanical methods used for cervical ripening [5]. The Foley catheter provides pressure at the internal cervical os, while the DBC, as a modified method, promotes cervical ripening by acting on both the internal and external cervix. Pharmacologic options include misoprostol (prostaglandin E1, PGE1) and dinoprostone (D, prostaglandin E2, PGE2). Both can increase the levels of prostaglandins (oral or intravaginal) in the body to induce uterine contractions. Both D vaginal gel and oral misoprostol, when used as a second cycle of induction after the failure of D vaginal insert, appear equally effective in achieving active labor and vaginal delivery, with no significant increase in adverse events [6]. Both mechanical and pharmacologic methods have been widely recognized for their safety and effectiveness in cervical ripening. However, there is still controversy about which cervical ripening option is better. Therefore, to compare the labor induction effects of the mechanical option and pharmacologic option, we performed a meta-analysis with DBC as the mechanical mode and D as the pharmacologic mode. Our study’s objective was to give medical professionals further information regarding IOL.
This study was previously registered with PROSPERO (CRD42024619409) and followed Preferred Reporting Items for Systematic Review and Meta-Analyses (PRISMA) guidelines. Following the rules set forth by PRISMA, we compiled a list of all published randomized controlled trials (RCTs) that evaluated cervical ripening effects by contrasting the D with the DBC. The following electronic databases were perused to March 2024: Cochrane libraries (https://www.cochranelibrary.com/), Sino Med (http://www.sinomed.ac.cn), PubMed (https://pubmed.ncbi.nlm.nih.gov/), and EMBAS (https://www.embase.com). Search terms included “Double-balloon catheter”, “Dinoprostone”, “Induction of labor”. Two independent researchers with expertise in obstetrics and gynecology systematically identified trials with full-text articles by screening the titles and abstracts, and discrepancies were resolved through discussion.
The selection criteria included: (1) RCTs published in English; (2) trials including women with live singleton pregnancies, cephalic presentation, intact fetal membranes, 3rd trimester pregnancies (
The exclusion criteria including: (1) unable to obtain full-text papers; (2) reviews, case reports, animal research or letters; (3) insufficient information for data extraction; (4) combined with other intervention, e.g., Foley catheter, misoprostol.
According to the selection and exclusion criteria, two independent reviewers with expertise in obstetrics and gynecology systematically identified trials with full-text articles. Discrepancies were resolved through discussion. The study selection process was illustrated in a PRISMA diagram.
Recorded data included in the study: author, year of publication, administration route and dose of D, balloon volume (mL), outcomes, sample number, patient age, outcome indicators, and baseline Bishop score. Data were managed and extracted by NoteExpress (version 3.20; Aegean Software, Beijing, China) and Excel (version 14.4.0; Microsoft Corporation, Redmond, WA, USA).
The rate of cesarean deliveries and the incidence of hospital admission prior to the commencement of active labor were the main results. The secondary outcomes included assisted vaginal delivery, failure to progress, failed induction, presence of nonreassuring fetal heart rate, neonatal asphyxia, improvement in Bishop score, neonatal intensive care unit admission (NICU), time to vaginal delivery, need for oxytocin administration, postpartum hemorrhage incidence, uterine hyperstimulation rate, and vaginal delivery within 24 h. Active labor was defined as cervical dilation of at least 4 cm accompanied by regular and effective uterine contractions. In cases of uterine hyperstimulation, management involved discontinuing the catheter or vaginal insert, or reducing the oxytocin infusion rate. Neonatal asphyxia was defined as an Apgar score [7] of 7 or lower at either 1 minute or 5 minutes after birth, accompanied by umbilical artery blood pH
According to the Cochrane risk of bias tool for randomised trials (RoB 2.0, Cochrane Collaboration, Oxford, United Kingdom), the following biases were evaluated: selection bias (random sequence generation), attrition bias (incomplete outcome data), performance bias (blinding of personnel and participants), reporting bias (selective reporting), detection bias (blinding of outcome assessment), and other sources of bias [8].
The data analysis program used was RevMan 5.4 (Cochrane Collaboration, Oxford, United Kingdom). The 95% confidence interval (95% CI) and odds ratio (OR) were used as the count data effect markers. The measurement data effect markers were the standard mean difference (SMD) or mean difference (MD) and the 95% CI. The heterogeneity was tested using the
The PRISMA flow diagram (Fig. 1) for this meta-analysis presents the detailed process of screening. The sum of 689 studies were found by initial screening. After duplicates were removed, a total of 408 studies were included. After the titles and abstracts were screened, 273 items were removed. A total of 8 studies were included as being RCTs. Other intervention, gestational age
Fig. 1. The Preferred Reporting Items for Systematic Review and Meta-Analyses (PRISMA) flow diagram. RCTs, randomized controlled trials.
The primary information of 7 pooled studies is presented in Table 1 (Ref. [9, 10, 11, 12, 13, 14, 15]). Every study was published in English and identified through electronic data search. A total of 2267 patients were included, 1127 of whom were treated with DBC, while 1140 were treated with D.
| Trial | Intervention | Sample size DBC/D | Age (years) | Baseline Bishop score | Outcomes | |||
| Double-balloon catheter (DBC) | Dinoprostone (D) | DBC | D | DBC | D | |||
| Cromi et al. 2012 [13] | Cook cervical ripener balloon (50 mL/50 mL) | Controlled-release vaginal insert (10 mg) | 105/103 | 34 (19–42)/33.6 | 33 (20–45)/32.9 | 2 (0–5)/2.05 | 2 (0–5)/2.06 | ①②③④⑤⑦⑧⑨⑩⑫⑬ |
| Du et al. 2015 [10] | Cook cervical ripener balloon (80 mL/80 mL) | Controlled-release vaginal insert (10 mg) | 76/79 | 28.45 | 27.33 | 3.38 | 3.76 | ①②③④⑤⑥⑦⑧⑨⑩⑪⑫⑬⑭ |
| Wang et al. 2014 [14] | Cook cervical ripener balloon (80 mL/80 mL) | Controlled-release vaginal insert (10 mg) | 67/59 | 27.9 | 27.8 | 2.4 | 2.5 | ①③⑤⑥⑦⑨⑩⑪⑫⑬⑭ |
| Suffecool et al. 2014 [12] | Cook cervical ripener balloon (80 mL/80 mL) | Controlled-release vaginal insert (10 mg) | 31/31 | 27.5 | 28 | 2 (0–5)/2.12 | 2 (1–5)/2.23 | ①②③④⑤⑧⑨⑬ |
| Shechter-Maor et al. 2015 [11] | Cook cervical ripener balloon (unclear) | Times-release formulation Propess (10 mg) | 26/26 | 28.5 (20–40)/28.89 | 28.5 (18–39)/28.5 | - | - | ①②⑤⑧⑩ |
| Beckmann et al. 2020 [9] | Cook cervical ripener balloon (80 mL/80 mL) | vaginal gel (2 mg) or controlled-release vaginal tape (10 mg) | 215/233 | 30 (26–34)/30 | 30 (26–34)/30 | 3 (2–4)/3 | 3 (2–4)/3 | ①②③④⑥ |
| Diguisto et al. 2021 [15] | Cook cervical ripener balloon (80 mL/80 mL) | Times-release formulation Propess (10 mg) | 607/609 | 31.1 | 31.3 | 3.2 | 3.1 | ①②③⑤⑥⑩⑪⑫⑬⑭ |
Note: data are displayed as number (%), median (range), or mean
The Jadad scale study indicated that low-quality literature had a score of
Fig. 2. Risk of bias graph. (A) A review of authors’ judgments about each risk of bias item presented as percentages across all included studies. (B) A review of authors’ judgments about each risk of bias item for each included study.
The primary outcomes were cesarean delivery rate and time to onset of active labor. All studies reported cesarean delivery rate, including 2267 patients (1127 in DBC group and 1140 in D group). The heterogeneity test yielded p = 0.63 and I2 = 0% as findings. Consequently, the overall effect size was computed using a fixed effects model. The information showed that there was no statistically significant difference involving the two groups [OR = 1.10, 95% CI: (0.91, 1.32), p = 0.34] (Fig. 3). Time to beginning of active labor was observed in 4 studies including 441 participants (222 in the DBC group and 219 in the D group). Heterogeneity test findings demonstrated that p
Fig. 3. Meta-analysis of cesarean delivery rate between double-balloon catheter (DBC) and dinoprostone (D). 95% CI, 95% confidence interval; M-H, Mantel-Haenszel.
Fig. 4. Meta-analysis of time to onset of active labor between DBC and D. IV, inverse variance; SD, standard deviation.
No statistically significant change was seen involving the 2 groups in the improvement to Bishop score [MD = –0.01, 95% CI: (–0.95, 0.94), p = 0.99] (Fig. 5). A sensitivity analysis based on the type of model showed consistent results [MD = 0.52, 95% CI: (–0.84, 1.88), p = 0.46, fixed model]. Furthermore, sensitivity analysis was conducted by removing studies. After excluding the study by Cromi et al. [13] on the improvement to Bishop score to avoid significant heterogeneity, there was a statistically significant variation [MD = 0.48, 95% CI: (0.12, 0.84), p = 0.010] involving the two groupings. However, compared to DBC group, the D group had a lower rate of oxytocin administration [OR = 3.96, 95% CI: (3.18, 4.93), p
Fig. 5. Meta-analysis of Bishop score between DBC and D.
Fig. 6. Meta-analysis of oxytocin administration requiring rate between DBC and D.
In contrast to the DBC group, the D group had a higher incidence rate of uterine hyperstimulation [OR = 0.16, 95% CI: (0.08, 0.32), p
Fig. 7. Meta-analysis of uterine hyperstimulation rate between DBC and D.
Fig. 8. Meta-analysis of neonatal asphyxia between DBC and D.
Fig. 9. Meta-analysis of the incidence of nonreassuring fetal heart a rate between DBC and D.
Fig. 10. Meta-analysis of NICU admission rate between DBC and D.
There was no statistically significant difference between the 2 groups in failure to progress rate [OR = 1.25, 95% CI: (0.75, 2.10), p = 0.40] (Fig. 11), failed induction rate [OR = 1.09, 95% CI: (0.54, 2.20), p = 0.80] (Fig. 12), time to vaginal delivery [MD = 0.33, 95% CI: (–3.16, 3.81), p = 0.85] (Fig. 13), vaginal delivery within 24 h [OR = 1.26, 95% CI: (0.53, 2.98), p = 0.60] (Fig. 14), assisted vaginal delivery rate [OR = 0.83, 95% CI: (0.65, 1.06), p = 0.13] (Fig. 15), or postpartum hemorrhage incidence rate [OR = 0.80, 95% CI: (0.60, 1.09), p = 0.16] (Fig. 16). A sensitivity analysis based on the type of model for failure to progress rate [OR = 1.04, 95% CI: (0.82, 1.32), p = 0.76, fixed model], time to vaginal delivery [OR = 0.71, 95% CI: (–0.67, 2.09), p = 0.31, fixed model], and vaginal delivery within 24 h [OR = 0.82, 95% CI: (0.64, 1.04), p = 0.11, fixed model] showed consistent results. Furthermore, sensitivity analysis was conducted by removing studies. After excluding the study by Du et al. [10] on the failure to progress rate to avoid significant heterogeneity, there was no statistically significant variation [OR = 0.98, 95% CI: (0.71, 1.35), p = 0.90] involving the two groupings. After excluding the study by Du and Wang et al. [10, 14] on the time to vaginal delivery to avoid significant heterogeneity, there was no statistically significant variation [MD = 0.05, 95% CI: (–1.67, 1.57), p = 0.95] involving the two groupings. After excluding the study by Cromi, Diguisto and Suffecool et al. [12, 13, 15] on the vaginal delivery within 24 h to avoid significant heterogeneity, there was no statistically significant variation [OR = 0.91, 95% CI: (0.57, 1.46), p = 0.69] involving the two groupings.
Fig. 11. Meta-analysis of failure to progress rate between DBC and D.
Fig. 12. Meta-analysis of failed induction rate between DBC and D.
Fig. 13. Meta-analysis of time to vaginal delivery between DBC and D.
Fig. 14. Meta-analysis of vaginal delivery within 24 h between DBC and D.
Fig. 15. Meta-analysis of assisted vaginal delivery rate between DBC and D.
Fig. 16. Meta-analysis of postpartum hemorrhage incidence between DBC and D.
The Begg’s funnel plots of both the primary and secondary outcomes indicated that no publication bias existed (Supplementary Figs. 1–14).
The findings of our meta-analysis comparing the labor induction effects of DBC and D provide valuable insights into the clinical management of labor induction. Overall, our analysis of 7 RCTs involving 2267 patients revealed no significant differences in primary outcomes such as cesarean delivery rate and time to onset of active labor between the DBC and D groups.
The process of cervical ripening including softening, effacement and dilatation is based on uterine contractile activity and immunologic effects [16]. By raising inflammatory mediators in the cervix and causing cervical remodeling, prostaglandins are essential for cervical ripening [17, 18]. D, one of the major prostaglandin formulations is approved in many countries worldwide for cervical ripening. It takes action on the prostaglandin type E prostanoid (EP) family of receptors [19]. Also, D can modulate the inflammatory response by increasing the inflammatory mediators such as interleukin-8 (IL-8) and tumor necrosis factor-a (TNF-a) in vivo [20]. The first balloon used to dilate the cervix was revealed by Chiari et al. in 1855 [21]. The direct mechanical pressure by the balloon can increase the release of prostaglandins from the decidua and cervix [22]. DBC is an improved version of the Foley catheter and is specifically designed for induction of labor. However, a meta-analysis reported that DBC and Foley catheter exhibit comparable efficacy, efficiency, safety, and patient satisfaction levels [5]. Both DBC and D are the most frequently used methods for cervical ripening. Under special circumstances, choosing between the two is not difficult. For example, the D vaginal insert group was associated with a higher rate of successful vaginal delivery within 24 hours as compared to DBC group in term nulliparous pregnancies, although it was also linked to a higher occurrence of nonreassuring fetal heart rate patterns [23]. However, for women without relevant contraindications, there is still much controversy over which method of cervical ripening is the best.
The results of our meta-analysis may provide some guidance. In this meta-analysis, we compared the labor induction effects between DBC and D. According to our research, there were no appreciable variations involving the 2 methods in most primary and secondary outcomes in terms of time to onset of active labor, success rate of induction, improvement to Bishop score, success rate of labor progress, time to vaginal delivery and vaginal delivery rate within 24 hours. The possible reason is that although D takes effect quickly, some patients may need an early removal due to the uterine hyperstimulation. Insufficient duration of drug action may lead to incomplete ripeness of the cervix. Our analysis revealed no discernible change in the time to onset of active labor between the DBC and D groups, suggesting that both methods are equally effective in initiating labor once the induction process has begun. Time to onset of active labor is a critical parameter in labor management, as delays in labor progression can increase the risk of maternal and neonatal complications [24]. Our findings support the notion that both DBC and D are viable options for achieving timely labor onset. Timely onset of active labor is crucial for minimizing maternal and neonatal complications associated with prolonged labor induction [24]. The Bishop score is a widely used assessment tool for predicting the success of labor induction and is based on cervical dilation, effacement, consistency, position, and fetal station [2]. Moreover, our results regarding no discernible change in the improvement of Bishop score involving the two groups indicated that both DBC and D can effectively improve cervical ripeness, leading to successful labor induction. However, we observed that the oxytocin administration rate was lower in the D group in contrast to the DBC group. This may be attributed to the rapid onset of action of D, leading to fewer cases requiring additional oxytocin augmentation. This finding suggests that D may require less additional intervention to augment labor progress, which could have implications for resource utilization and patient satisfaction.
In terms of side effects, there was not a noticeable variation in the cesarean delivery rate, postpartum hemorrhage incidence rate, or assisted vaginal delivery rate between DBC and D. In our meta-analysis, we observed no discernible change in the cesarean delivery rate between the DBC and D groups. This result is in line with earlier researchers that have reported comparable rates of cesarean delivery following labor induction with mechanical and pharmacological methods [25]. Cesarean delivery rates are important outcomes in labor induction, as they reflect both maternal and fetal outcomes and are influenced by various factors, including the method of induction, maternal characteristics, and obstetric practices [26]. Compared to DBC, D was more likely to cause uterine hyperstimulation, leading to a higher rate of nonreasuring fetal heart and neonatal asphyxia. These findings are consistent with the pharmacological properties of D, which can induce strong and sometimes unpredictable uterine contractions [27]. Clinicians should weigh the potential benefits of reduced oxytocin use against the risks of uterine hyperstimulation when choosing between DBC or D for labor induction. However, there was not a noticeable variation in neonatal intensive care unit (ICU) admission rates involving the two groups. A possible reason is that neonatal asphyxia caused by uterine hyperstimulation is reversible [28]. Immediate conservative measures (remove medication or use of tocolytic agents) can be undertaken to reverse the uterine hyperstimulation state. The results suggest that when using D for cervical ripening, it is necessary to closely monitor uterine contractions and the changes of fetal heart rate. If regular contractions have already occurred, we suggest avoiding using D.
Our study has several strengths, including the inclusion of a large sample size from multiple RCTs and a comprehensive analysis of various labor induction outcomes. However, there are some limitations to consider in this study. Firstly, due to the obvious difference between the 2 interventions, blinding for participants or researchers cannot be implemented. The blinding of participants and personnel in all trials was judged as unclear risk. Secondly, the comparison of pain perception between the 2 methods by patients has not been revealed. Moreover, the included studies enrolled diverse patient populations, with some studies specifically focusing on women with prolonged gestation, hypertension, diabetes, oligohydramnios, IUGR, and/or suspected fetal macrosomia. However, the level of detail regarding these conditions varied, and some studies did not provide comprehensive information on all comorbidities. As a result, a precise classification of the populations could not be achieved, but we have provided a general description of the most common conditions observed across the studies. Therefore, more high-quality studies are needed for a future meta-analysis. Finally, the sensitivity analysis conducted in this study helped clarify the sources of heterogeneity, especially in outcomes like time to onset of active labor and improvement to Bishop score. These analyses provided more consistent and reliable conclusions by excluding studies that contributed to significant heterogeneity. However, for outcomes such as nonreassuring fetal heart rate and vaginal delivery within 24 hours, heterogeneity remained unaddressed, suggesting that future studies should focus on standardizing reporting methods to minimize variation.
DBC was equally effective as D in terms of success rate of induction, time to onset active labor, success rate of induction, improvement to Bishop score, success rate of progress, time to vaginal delivery, vaginal delivery rate within 24 hours, cesarean delivery rate, postpartum hemorrhage incidence rate, and assisted vaginal delivery rate, and incidence of nonreasuring fetal heart rate. Although the oxytocin administration rate was lower, D was more likely to cause uterine hyperstimulation, leading to a higher rate of neonatal asphyxia. Clinicians should consider individual patient factors and preferences when selecting the appropriate method for cervical ripening.
The datasets used and analyzed during the current study are available from the corresponding author on reasonable request.
XZ was responsible for conceptualization, data curation, formal analysis, investigation, methodology, project administration, resources, software, supervision, validation, and drafting the original manuscript. YW contributed to conceptualization, data curation, formal analysis, investigation, methodology, resources, software, supervision, validation, and funding acquisition, as well as reviewing and editing the manuscript. Both authors have read and approved the final manuscript. Both authors have participated sufficiently in the work and agreed to be accountable for all aspects of the work.
Not applicable.
Not applicable.
This research received no external funding.
The authors declare no conflict of interest.
Supplementary material associated with this article can be found, in the online version, at https://doi.org/10.31083/CEOG26977.
References
Publisher’s Note: IMR Press stays neutral with regard to jurisdictional claims in published maps and institutional affiliations.