The search strategy was designed to capture all relevant studies evaluating labor induction with misoprostol in twin pregnancies. Keywords were selected based on MeSH terms and text words identified in prior related reviews, and included: “twin pregnancy”, “multiple pregnancy”, “labor induction”, “misoprostol”, “prostaglandin”, and “perinatal outcome”.

Distinct search strategies were conducted across the following databases: COCHRANE, EMBASE, MEDLINE, and CINAHL, covering the period from January 1, 1990, to December 12, 2023. Only studies published in English or French were included in this literature review. Based on similar searches in the literature and comparable systematic reviews, the following search terms were used: Twin pregnancies, Twin pregnancy, Twin, Twins, Sibling, Siblings, Reciprocal, Reciprocals, Induced labor, Induction of labor, Induction labor, Inductions labor, Induce labor, Labor induced, Labor induction, Labor inductions, Induction of labour, Inducing labor, Induced labour, Induce labour, Inductions labour, Labour induced, Labour induction, Labour inductions, Inducing labour, Induce Delivery, Misoprostol, Misoprostol induction, Prostaglandin, Gestation pregnancies, Obstetric labor, Expectant, Single pregnancy, Pregnancy, Pregnancies, Ocytocin, Syntocinon, Natural childbirth.

These keywords, searched in titles and abstracts, were combined using the Boolean operators “AND” and “OR” following the PICOT method. Additional articles were identified from the reference lists of selected studies and relevant systematic reviews. Unpublished randomized controlled trials (RCTs) were also searched on the clinical trials website ( http://clinicaltrials.gov ) using the keywords twin, multiple pregnancy, and labor induction.

Randomized controlled trials and observational studies (cohort, cross-sectional, and case-control studies) were included if they compared neonatal outcomes based on the mode of labor induction in twin pregnancies. The primary neonatal outcomes were the 5-minute Apgar score and neonatal mortality. Twin pregnancies could be either dichorionic or monochorionic-diamniotic. The gestational age had to be at least 32 weeks, as twins born before this period may present complications unrelated to the mode of delivery. Both Twin A and Twin B had to weigh at least 1.5 kg.

Studies were excluded if they focused exclusively on singleton pregnancies, included non-viable fetuses, had a gestational age below 34 weeks, or involved women with previous uterine surgery or known uterine anomalies. This approach ensured the review focused on the most clinically relevant and comparable studies.

Once the article search was completed, the references were imported into Filemaker Pro 2023 and then EndNote. Duplicates were removed, and the titles and abstracts of the remaining studies were screened to exclude those irrelevant to the research question. The full texts of potentially relevant studies were reviewed to assess their eligibility based on predefined criteria.

A specific data extraction form was designed using Filemaker 23 to standardize the collection of information and ensure consistency. The form recorded the study type, authors’ names, title, year of publication, study design, country, objectives, results, inclusion and exclusion criteria, number of patients, gestational age of twins, presentation of the second twin, mode of delivery for each twin, birth weight, number of twins with an Apgar score < 7 at 5 minutes, neonatal deaths by twin and delivery mode (vaginal or cesarean), and maternal outcomes.

The extracted data were recorded in customized tables.

The selected studies were critically evaluated using the Cochrane Collaboration tool, which assesses the risk of bias as high, low, or unclear across five domains: selection, performance, attrition, reporting, and other biases [15] . Criteria included randomization methods, blinding processes, handling of exclusions, and other potential biases.

Additionally, the included studies were assessed using the GRADE (Grades of Recommendation, Assessment, Development, and Evaluation) approach [16] [17] . The GRADE system assigns a level of evidence (high, moderate, low, or very low) to each outcome based on the following criteria: High: Randomized trials or observational studies with two levels of upgrade; Moderate: Downgraded randomized trials or upgraded observational studies; Low: Randomized trials with two levels of downgrade or observational studies; Very low: Randomized trials with three levels of downgrade, downgraded observational studies, or case series/reports.

Based on these assessments, the GRADE system provides either a strong or weak recommendation for an intervention.

The methodology, interventions, and outcomes of each included study were described and grouped into logical categories. After analyzing and grouping the studies, their results were synthesized to produce a summary, taking into account differences that might influence their conclusions.

A meta-analysis could not be performed due to several limiting factors. First, the collected data were not sufficiently homogeneous to allow a meaningful combined analysis. Second, the reporting methods varied significantly across studies, making direct comparisons and synthesis challenging. Third, the labor induction methods compared to misoprostol were not uniformly represented in the studies. Finally, the small number of available studies (only five) limited the statistical power and generalizability of any conclusions that could have been drawn from a meta-analysis.

Table 1 summarizes the risk of bias and GRADE assessment for each of the five included studies.

Each study is assigned a risk of bias rating (low, high, or unclear) for each domain. Additionally, the GRADE score (high, moderate, low, or very low) is provided for the overall quality of evidence.

RC: Restrospective cohort

The five selected studies [13] [14] [18] - [20] aimed to evaluate the efficacy and safety of labor induction in twin pregnancies, as well as neonatal and maternal outcomes. Some studies compared the use of misoprostol for labor induction in twin pregnancies to spontaneous delivery [20] , oxytocin [14] , or dinoprostone [18] . One study focused on the association between labor induction and the risk of cesarean delivery [13] .

The five studies included a total of 1205 participants. Jonsson et al. (2015) [13] had the largest sample size with 462 participants, followed by Mikaelsen et al. (2022) [20] with 354 participants. In contrast, Simoes et al. (2006) [19] had the smallest sample size with 69 participants.

In all studies, participants had a gestational age of at least 34 weeks, except for Simoes et al., where it was 35 weeks or more. Exclusions included intrauterine fetal death of one or both twins, congenital anomalies incompatible with life, and gestational age below 34 weeks.

Bush et al., Huber et al., and Simoes et al. excluded pregnancies where the first twin was non-cephalic. Simoes et al. also excluded multiparous women, those with a uterine scar, those opting for cesarean delivery, and those with contraindications to labor induction. Jonsson et al. (2015) excluded women who delivered the second twin via cesarean after a vaginal delivery of the first twin and those with a uterine scar.

Regarding the results, the studies by Mikaelsen et al. and Jonsson et al. present them based on whether the labor was spontaneous or induced. The studies by Huber et al. and Bush et al. present their data according to the molecule used for induction (oral misoprostol/vaginal dinoprostone for Huber et al. or misoprostol/oxytocin alone for Bush et al.), while Simões et al. present them by comparing labor induction and elective cesarean section.

In the following Table 2 , the socio-demographic characteristics of the women in the five studies are summarized.

Across the five studies examined, the age of the mothers showed no significant difference, whether they underwent labor induction or began labor spontaneously. Similarly, no notable differences were observed among women who received misoprostol, dinoprostone, or oxytocin for induction, as well as those who delivered via cesarean section.

According to the study by Mikaelsen et al., the group of patients whose labor was induced included a higher number of nulliparous women. A similar observation was made in the study by Bush et al., where the group receiving misoprostol for labor induction had more nulliparous women than the group receiving oxytocin alone.

In the studies conducted by Jonsson et al. and Huber et al., parity did not show a significant difference between the studied groups, namely induced labor versus spontaneous labor, and oral misoprostol compared to vaginal dinoprostone. However, the study by Simões et al. did not specify parity. In this latter study, all women were nulliparous.

Only three of the five studies (Simoes et al., 2006; Bush et al., 2006; and Huber et al., 2014) addressed the issue of maternal morbidity, which was not evaluated in the studies by Jonsson et al., 2015 and Mikaelsen et al., 2022. These three studies also examined neonatal morbidity, assessed by umbilical artery pH in newborns (except in the study by Simoes et al., 2006), Apgar score < 7, admission to the neonatal intensive care unit, low birth weight, the possibility of trauma, and the presence of meconium-stained amniotic fluid before labor.

As for maternal morbidity, it was evaluated based on the occurrence of postpartum hemorrhage and the amount of blood lost, the incidence of postpartum infection, uterine rupture, and the performance of a hysterectomy.

In the studies by Jonsson et al., 2015 and Mikaelsen et al., 2022, labor was initiated either spontaneously or by amniotomy, oxytocin infusion, cervical ripening with a Foley catheter, administration of prostaglandins, or intravaginal administration of a catheter balloon, misoprostol, or dinoprostone based on the Bishop score. In the study by Simoes, induction was performed either by administration of misoprostol alone or in combination with oxytocin. Similarly, the study by Bush used either intravaginal misoprostol alone or oxytocin alone. The study by Huber, on the other hand, induced labor by administering oral misoprostol or intravaginal dinoprostone.

Finally, the study by Huber ensured patient monitoring by cardiotocography for at least 60 minutes after the respective administration of different prostaglandins to detect potential fetal distress or uterine hyperstimulation.

The following Table 3 shows the characteristics of the participants in the five studies.

Mikaelsen et al. and Jonsson et al. presented results based on whether labor was spontaneous or induced. Huber et al. and Bush et al. compared outcomes based on the induction method (oral misoprostol/vaginal dinoprostone for Huber et al.; misoprostol/oxytocin alone for Bush et al.). Simoes et al. compared labor induction to elective cesarean delivery ( Table 4 ).

All studies except Huber et al. (2014) evaluated the Apgar score at 5 minutes after birth. Huber et al. (2014) and Bush et al. (2006) measured the Apgar score at 1 minute.

In the study conducted by Huber et al. (2014), no significant differences were observed in the 1-minute Apgar score between the groups receiving misoprostol and dinoprostone for labor induction. Additionally, no notable differences were found between the twins or based on the mode of delivery, whether vaginal or cesarean.

Mikaelsen et al. (2022) and Jonsson et al. (2015) reported that a 5-minute Apgar score < 7 was more common in the second twin. In Mikaelsen et al., induced labor was associated with a higher frequency of a 5-minute Apgar score < 7 in the first twin. However, in Jonsson et al., the mode of labor induction did not appear to have any impact on the Apgar score.

In the study by Simoes et al. (2006), only two newborns in the cesarean group had a 5-minute Apgar score < 7, indicating that low Apgar scores were rare in this group.

Finally, Bush et al. (2006) found no significant differences in the Apgar scores at either 1 or 5 minutes between the groups receiving misoprostol and oxytocin for labor induction.

Overall, these findings suggest that while there are some variations in Apgar scores between twins and induction methods, the differences are not consistently significant across studies. This highlights the need for further research to better understand the factors influencing neonatal outcomes in twin pregnancies.

In the study by Huber et al. (2014), no significant differences were found in the mean umbilical artery pH between the groups receiving misoprostol and dinoprostone for labor induction. Additionally, no differences were observed between the twins or based on the mode of delivery, whether vaginal or cesarean.

Mikaelsen et al. (2022) reported that the mean umbilical artery pH was lower in the second twin in both the spontaneous and induced labor groups. However, the frequency of a pH < 7.0, which is indicative of metabolic acidosis, was low and did not show significant differences between the groups.

Similarly, Bush et al. (2006) and Jonsson et al. (2015) found no significant differences in the mean umbilical artery pH between the groups studied, suggesting that the choice of induction method did not significantly impact this neonatal outcome.

Regarding Neonatal Intensive Care Unit (NICU) admission, Bush et al. (2006) reported no significant differences in NICU admission rates between the groups receiving misoprostol and oxytocin for labor induction. In contrast, Simoes et al. (2006) documented one case of NICU admission due to respiratory difficulties in each group—both in the induced labor group and the cesarean group.

These findings indicate that while there are some variations in umbilical artery pH and NICU admission rates, the differences are not consistently significant across studies. This underscores the need for further research to better understand the factors influencing neonatal outcomes in twin pregnancies, particularly in relation to different labor induction methods.

Where available, outcomes were also examined based on chorionicity. Two studies (Huber et al. and Mikaelsen et al.) reported data stratified by monochorionic and dichorionic twins, though detailed subgroup analyses were limited. No significant differences in maternal morbidity or neonatal outcomes were observed between chorionicity groups. Neonatal complications reported across studies included respiratory distress syndrome (3% - 8%), NICU admission rates (20% - 35%), and low Apgar scores at 5 minutes (<7) in approximately 5% - 10% of neonates. Umbilical artery pH values were available in three studies and did not differ significantly between groups.

Maternal morbidity was evaluated in the studies conducted by Simoes et al. (2006), Huber et al. (2014), and Bush et al. (2006). Notably, no cases of uterine rupture were reported in any of these studies.

In the study by Simoes et al. (2006), there were no reported cases of uterine hyperstimulation. However, one case of postpartum hemorrhage and one case of postpartum infection were identified in the cesarean group.

Bush et al. (2006) reported one case of postpartum hemorrhage in the oxytocin group, which was successfully managed medically without the need for blood transfusion. Importantly, no hysterectomies or severe maternal complications were observed in this study.

In the study by Huber et al. (2014), labor induction was associated with certain risks, but a lower risk of postpartum hemorrhage was observed in the misoprostol group compared to the dinoprostone group. However, these differences did not reach statistical significance, indicating that while there may be a trend, it was not strong enough to draw definitive conclusions.

Overall, these findings suggest that while maternal morbidity risks, such as postpartum hemorrhage and infection, are present in labor induction for twin pregnancies, they are generally manageable and do not frequently result in severe complications. The choice of induction agent may influence these outcomes, but further research is needed to clarify these associations.

The results of the studies showed that labor induction in twin pregnancies was not associated with significant differences in maternal morbidity, regardless of the induction agent used. No notable differences were observed in cesarean rates between spontaneous and induced labor groups. Similarly, neonatal outcomes, such as low Apgar scores at 5 minutes and umbilical artery pH values, did not show significant differences between spontaneous and induced labor groups, although some studies reported a higher frequency of Apgar scores < 7 at 5 minutes in the second twin.

The findings of this systematic review are consistent with previously published evidence on labor induction in twin pregnancies. The included studies provided important insights into the use of different labor induction methods, such as misoprostol, dinoprostone, and oxytocin, and compared their effects on maternal and neonatal outcomes.

In the context of labor induction for twin pregnancies, recent findings on singleton pregnancies with an unfavorable cervix are relevant. A systematic review and meta-analysis up to July 2013 found that intravaginal misoprostol was potentially more effective than intracervical dinoprostone for women with a Bishop score <6. Misoprostol increased the rate of vaginal deliveries within 24 hours and required less oxytocin augmentation compared to dinoprostone [21] . However, misoprostol was associated with a higher incidence of uterine hyperstimulation and tachysystole. Cesarean rates, NICU admissions, and Apgar scores at 1 and 5 minutes did not differ significantly between the two groups.

While intravaginal misoprostol appears more effective for labor induction in singleton pregnancies, dinoprostone offers a safety advantage due to its lower association with uterine hyperstimulation and tachysystole. For twin pregnancies, the direct implications of these findings remain to be studied, highlighting the need for further research on the efficacy and safety of misoprostol and dinoprostone specifically for twin pregnancies.

Access to medications and their cost are crucial factors influencing clinical practices, especially in resource-limited settings. Dinoprostone, commonly used in developed countries for labor induction, is often inaccessible in Africa due to supply constraints and high costs. In contrast, misoprostol, which is cheaper and more readily available, has become an attractive alternative for many healthcare providers in Africa. Its stability at room temperature, ease of storage, and lower cost make it particularly suitable for African conditions. However, adequate training for healthcare providers on the correct use of misoprostol is essential to minimize risks and ensure safe and effective care.

A 2021 Cochrane review by Kerr et al. focused on the use of low-dose oral misoprostol for labor induction [22] . This review highlighted that low-dose oral misoprostol (≤50 µg) is effective and safe for labor induction in the third trimester, with a reduction in cesarean rates and uterine hyperstimulation compared to vaginal dinoprostone and misoprostol. However, it may slightly prolong the time to delivery. These findings are consistent with some conclusions of this review, particularly regarding the efficacy of misoprostol compared to other induction methods.

Given these findings, this meta-analysis reaffirms some previous evidence and offers a nuanced perspective on others. The main conclusion is that induction, particularly with misoprostol, can be a safe and effective option for twin pregnancies, but the decision must always consider the individual clinical context and patient preferences. Existing controversies in the literature underscore the need for further research, multicenter studies, and randomized clinical trials to provide clear guidelines for clinicians.

In conclusion, this systematic review provides additional evidence on the use of labor induction in twin pregnancies. The results indicate that labor induction in these women does not appear to increase the risk of significant maternal or neonatal complications. However, the included studies show variations in induction protocols and evaluated criteria, which may influence the results. Therefore, decisions regarding labor induction in twin pregnancies should be made on a case-by-case basis, considering individual factors and current clinical recommendations. Further research could help clarify outcomes and provide more specific guidelines for clinical practice.

The findings of the five studies included in this review hold significant value for practitioners involved in the management of twin pregnancies and labor induction. These insights can help refine clinical decision-making and improve patient care.

One of the key takeaways is the importance of selecting an appropriate induction method. The studies by Huber et al. [18] and Bush et al. [14] revealed no substantial differences in efficacy and safety between misoprostol and dinoprostone. This knowledge allows practitioners to base their choice of induction method on individual patient characteristics, optimizing both maternal and fetal outcomes.

Maternal morbidity is another critical concern. Three of the studies focused on maternal complications such as postpartum hemorrhage, infections, and uterine rupture. Given these risks, practitioners must maintain heightened vigilance throughout labor induction in twin pregnancies, ensuring that proper monitoring and management protocols are in place to safeguard maternal health.

Furthermore, twin pregnancies are inherently associated with an increased risk of preterm birth, leading to potential neonatal complications. The review underscores the necessity for practitioners to anticipate challenges such as NICU admissions and low Apgar scores. Being prepared to provide immediate neonatal intensive care when required can significantly enhance neonatal outcomes.

Despite the insights gained, this systematic review has also illuminated significant gaps in research regarding labor induction in twin pregnancies. Addressing these gaps is essential for advancing clinical knowledge and improving patient outcomes.

One major limitation is the scarcity of twin-specific studies. Many of the reviewed studies were not designed exclusively for twin pregnancies, often providing insufficient data on outcomes for each twin individually. Future research should focus explicitly on twin pregnancies, allowing for a more nuanced understanding of how labor induction affects each twin separately.

The quality of available studies also varies considerably, with some exhibiting a high or uncertain risk of bias. To strengthen the evidence base, well-structured randomized controlled trials are needed. These studies should rigorously evaluate the safety and effectiveness of different labor induction methods, ensuring that clinical decisions are grounded in high-quality data.

Moreover, outcome evaluation in existing studies remains inconsistent. Some studies fail to comprehensively assess crucial aspects such as maternal morbidity or neonatal complications specific to twin pregnancies. Future research should incorporate thorough and standardized evaluations of both maternal and neonatal outcomes to provide clearer guidance for clinical practice.

Beyond the clinical setting, the findings of this review also have broader public health implications, particularly concerning policies and education surrounding the management of twin pregnancies and labor induction.

A key public health priority is ensuring that practitioners receive adequate training. Clinicians managing twin pregnancies must be well-versed in various induction methods, associated risks, and best practices for handling maternal and neonatal complications. Structured training programs can enhance practitioners’ competencies, leading to improved patient outcomes.

Health promotion initiatives also play a vital role in optimizing care for twin pregnancies. Educating expectant mothers about the medical necessity and benefits of labor induction, when applicable, empowers them to make informed decisions in partnership with their healthcare providers. Increased awareness can lead to better adherence to medical recommendations and, ultimately, improved perinatal outcomes.

By addressing these clinical, research, and public health implications, healthcare providers and policymakers can work together to refine protocols for labor induction in twin pregnancies, ensuring safer and more effective care for mothers and newborns alike

The studies included in this systematic review have several strengths. They were conducted following Cochrane guidelines, ensuring a robust and systematic methodology. The objective of the systematic review was clearly defined, with pre-established inclusion and exclusion criteria regarding the patient population, interventions, comparators, outcomes, and study design. A comprehensive and systematic literature search was performed across multiple databases, supplemented by a manual search of reference lists from retrieved articles. Additionally, data extraction and validation were conducted independently by different researchers, enhancing the reliability of the findings.

However, these studies also have certain limitations. The potential for data pooling in a meta-analysis was restricted due to heterogeneity in study content, intervention duration, and outcome measures. As a result, the conclusions of this review are primarily based on a narrative synthesis, which may limit the generalizability of the findings.

The strength of this systematic review lies in its explicit and robust methodology, which ensures transparency, accuracy, and reproducibility, reducing the risk of bias. However, the review has limitations, including heterogeneity between studies, which prevented a meta-analysis, and its retrospective nature, which may introduce bias.

The conflicting findings across studies may be attributed to several factors. First, methodological limitations, such as retrospective design and lack of randomization, may introduce bias. Additionally, differences in patient populations, including variation in parity, gestational age at induction, and chorionicity, can influence outcomes. Lastly, inconsistency in induction protocols (e.g., dosage, route of administration, monitoring procedures) may further explain heterogeneity in the results. These variations underscore the importance of standardizing study designs and clearly reporting subgroup characteristics in future research.

Systematic reviews must be continuously updated to ensure their relevance and accuracy.