External cephalic version in nonvertex second twin—Success rate, mode of delivery, and safety: A systematic review

Abstract

Introduction

One of the key challenges regarding the management of twins involves choosing the optimal mode of delivery, which is strongly influenced by the final presentation of both fetuses. In cases of vertex–nonvertex pregnancies attempting the trial of vaginal delivery, external cephalic version (ECV) is one of possible management options. The main objective of this review was to collect and summarize available data in terms of the application of ECV in the population of nonvertex second twins.

Material and Methods

Using the PRISMA guidelines, we searched for original, English‐language studies investigating ECV in nonvertex second twins. The PubMed/MEDLINE, SCOPUS, and COCHRANE databases were searched until May 2024. Reviews, case reports, editorials, and conference papers were excluded from further analysis. Out of 260 papers retrieved, 10 were subjected to the final analysis in terms of success rates, modes of delivery, and adverse outcomes.

Results

The total number of ECVs was 289, with an overall success rate of 64.4%. In the group of successful versions, vertex vaginal delivery was achieved in 171 cases (91.9%). The incidence of adverse maternal and neonatal outcomes was low.

Conclusions

The purpose of this review was to consolidate and update the current knowledge regarding ECV in nonvertex second twins. Based on the results of this series of studies, ECV appears to be a reasonable management option. However, it is important to highlight several significant limitations. The primary concern is the lack of recent research in this field over the past three decades, with the most recent study in our review being published in 1998. Furthermore, the actual number of studies addressing this topic is relatively low, characterized by a retrospective nature and questionable methodologies. These limitations make it challenging to draw definitive conclusions for clinical practice. This is an important message for our community, emphasizing the need for further studies in this area, particularly randomized controlled trials, to evaluate the safety and success rate of vaginal twin delivery after ECV when the second twin presents in a nonvertex position.

This review provides updated knowledge on the external cephalic version in non‐vertex second twins. Although the results suggest that the procedure seems reasonable, the low quality of the studies, and more importantly, the lack of recent research, are significant limitations. The key message is the urgent need for new, high‐quality research in this area, particularly to evaluate the safety and success rates of this procedure in clinical practice.

Abbreviation

ECV

external cephalic version

Key message.

This review suggests external cephalic version for nonvertex second twins may be reasonable, but the lack of recent, high‐quality studies limits strong, reliable conclusions. New randomized trials are essential to better evaluate its safety and effectiveness.

1. INTRODUCTION

In recent decades, the number of twin pregnancies has significantly increased. ¹ , ² The selection of the optimal mode of delivery constitutes one of the key problems in the management of such pregnancies, which is strongly influenced by the final presentation of both fetuses. The problem becomes even more challenging in cases of the vertex–nonvertex configuration. Several retrospective analyses showed that planned cesarean delivery reduced the rate of complications in such a situation. ³ , ⁴ However, data from the randomized Twin Birth Study revealed no difference in adverse neonatal outcomes between planned cesarean delivery and the vaginal delivery group. ⁵

The conclusions of the Term Breech Trial, which investigated perinatal outcomes in singleton pregnancies with a breech presentation in relation to the mode of delivery, demonstrated higher rates of adverse neonatal outcomes in the vaginal delivery group compared to planned cesarean sections. ⁶ Despite various critical opinions regarding the methodology of this trial in particular, its results were incorporated into numerous obstetric guidelines, leading to a dramatic drop in the rate of vaginal breech deliveries in singleton pregnancies. ⁷ , ⁸ We believe that the recommendations may influence perinatal care in multiple pregnancies because the lack of proper training and experience in vaginal breech deliveries in singletons also contributes to abandoning the trial of vaginal delivery in vertex–nonvertex twins.

In cases of malpresenting singletons near term, an external cephalic version (ECV) is often offered. ⁹ , ¹⁰ The successful maneuver enables the trial of vaginal delivery in a vertex presentation. Although success rates vary widely, the procedure is known to be safe, with a very low risk of complications. ¹¹ ECV may also be performed in twins, mostly as an intrapartum procedure, in cases of the nonvertex presentation of the second twin after the delivery of the first one.

Available literature, guidelines, and textbooks predominantly focus on the application of ECV in singleton pregnancies, with very limited studies addressing its use in twins. This systematic review aims to compile data from available sources and investigate the efficacy, mode of delivery, and safety of intrapartum ECV performed in the nonvertex second twin after the delivery of the first one. To the best of our knowledge, this is the only systematic review concerning this subject.

2. MATERIAL AND METHODS

The review was developed according to the Updated Preferred Reporting Items for Systematic Reviews and Meta‐Analyses Statement (PRISMA) 2020. ¹² We used the following databases as sources: the MEDLINE/PubMed, Scopus, and the Cochrane Library. The last search was performed on the 8th of May 2024, and the detailed search strategy is presented in Appendix S1.

We retrieved 260 papers. Using the automatic search function in EndNote X9 (Clarivate Analytics, London, UK), 41 studies were deleted and further 10 studies were excluded manually. The remaining 209 papers were screened by two authors (M.D. and M.Z.). Non‐English language papers, case reports, conference papers, reviews, guidelines, and surveys were excluded. Studies concerning singleton pregnancies, maneuvers other than ECV, or lacking information on the performed ECV were also eliminated.

Subsequently, complete manuscripts were reviewed in detail by two authors (M.D., M.Z.). To present a more comprehensive analysis, it was essential not only to present success rates of the procedure but also to address the events that followed. Therefore, we considered a study as eligible only if it met at least 3 out of 4 following criteria: number of ECVs performed, number of successful ECVs, number of vertex vaginal deliveries after a successful procedure, and adverse outcomes. In order to avoid omitting any study relevant to the topic, we reviewed the papers that cited the identified manuscripts using Google Scholar, and we found an additional manuscript that was included in our series. ¹³ Disagreements at any step were clarified through discussion among all authors. Ultimately, 10 papers were included in this review. The entire selection process is presented in supplementary material—Figure S2.

The risk of bias was assessed independently by two authors (M.Z. and M.D.), using the National Institutes of Health (NIH) Quality Assessment Tool for Observational Cohort and Cross‐Sectional Studies, which was modified for the purpose of this review. ¹⁴ Answers “yes,” “no,” and “cannot determine/not applicable” were provided for each item included in the tool. A potential risk of bias was identified in cases when “no” or “cannot determine” were selected. Items awarded with “not applicable” were excluded from further assessment. Based on the rating guidelines provided with the assessment tool, each study was assigned to one of the three categories: “good” for studies with the lowest risk of bias, “fair” for studies susceptible to some risk of bias, and “poor” for studies representing a high risk of bias. If particular ratings for a given study differed between two assessing authors, the final conclusion was made through consensus with all authors. The results are presented in supplementary material—Table S4.

3. RESULTS

3.1. Main characteristics and risk of bias

We retrieved 10 papers, which met the inclusion criteria and were eligible for further analysis. ¹³ , ¹⁵ , ¹⁶ , ¹⁷ , ¹⁸ , ¹⁹ , ²⁰ , ²¹ , ²² , ²³ Retrospective papers were the most common type. Eight studies came from the United States of America, one from Israel, and one from Canada. The oldest paper was published in 1983, and the latest came from 1998. All four previously established inclusion criteria were met by six studies. ¹⁶ , ¹⁷ , ²⁰ , ²¹ , ²² , ²³

The total number of patients with vertex–nonvertex pregnancies was 1094. A study with the largest group was published by Greig et al. (208 vertex–nonvertex pregnancies). ¹⁹ The total number of ECVs was 289, with the highest number being reported by Kaplan (96 attempts). ¹⁸ The authors of three studies reported on chorionicity and amnionicity. ¹⁵ , ¹⁶ , ¹⁹ Eight authors referred to the estimated weight of fetuses included in their cohort. ¹³ , ¹⁶ , ¹⁸ , ¹⁹ , ²⁰ , ²¹ , ²² , ²³ The main characteristics of all included studies are presented in supplementary material—Table S3.

Five studies were categorized as “poor,” indicating a high risk of bias. One study was classified as “good,” representing the lowest risk, and the remaining four were marked as “fair,” indicating a moderate risk. The lack of clear and consistent definitions of particular outcomes, absence of control for confounding variables, and low participation rates were identified as the most common factors responsible for a high risk of bias. A detailed evaluation of each study is presented in supplementary material S2—Table S4.

3.2. Success rate and mode of delivery

Data published by Tchabo et al. regarding the success rate and mode of delivery were inconsistent. ¹⁵ Although the authors reported on the total number of 30 ECV attempts, with 27 of them resulting in vertex vaginal delivery, no precise information was provided about the total number of successful procedures, irrespective of the subsequent mode of delivery. As no contact details were included in the original manuscript, it was impossible to clarify our doubts with the author. For this reason, we decided to exclude this paper from our analysis of success rates and modes of delivery.

The number of successful ECVs, defined as obtaining the cephalic presentation of the nonvertex second twin after the delivery of the first one, was 186, representing 64.4% of all 289 attempts. Individual success rates varied between studies, ranging from 42% to 80%. Multiparous women were dominant in studies regarding maternal parity. ¹⁶ , ²⁰ , ²³ Breech was the most common presentation in three studies reporting on the type of nonvertex presentation before version. ¹⁶ , ²¹ , ²² Two authors used tocolysis and five applied epidurals during ECV. ¹⁶ , ¹⁸ , ²⁰ , ²¹ , ²² In five studies, oxytocin or artificial rupture of membranes were performed after the successful version. ¹⁶ , ¹⁸ , ²⁰ , ²¹ , ²²

A total of 171 fetuses were delivered in a vaginal cephalic presentation, accounting for 91.9% of all 186 successful attempts. A cesarean section was performed in the remaining 15 cases, with cord prolapse and fetal distress being the most common indications.

The group of unsuccessful procedures included 103 patients, accounting for 35.6% of all ECV attempts. Regarding the mode of delivery, there were 52 (50.5%) cesarean sections and 51 (49.5%) breech deliveries.

The results concerning success rates and the mode of delivery are shown in Figure 1. More details are presented in Table 1.

FIGURE 1.

Success rate and mode of delivery.

TABLE 1.

Success rate and mode of delivery.

	Authors and year	Mean maternal age (weeks)	Parity	Mean gestational age (weeks)	Mean birth weight of the second twin (grams)	Use of epidural during ECV	Use of tocolysis during ECV	Number of ECVs	Type of nonvertex presentation of the second twin	Number of successful ECVs (% of total attempts)	Vaginal vertex delivery after successful ECV (% of successful procedures)	Use of oxytocin or rupture of membranes as augmentation after successful ECV	CC after successful ECV	Indications for CC after successful ECV	Unsuccessful ECV (% of total attempts)	Mode of delivery after unsuccessful ECV
1.	Chervenak at el. 1983 21	a	a	a	a	Yes	a	25	Breech—11 Transverse—14	20 (80%)	18 (90%)	Yes	2	Failure to descend—1 Twin A placental obstruction—1	5 (20%)	CC—3 Breech extraction—2
2.	Gocke et al. 1989 20	28.0	Nullipara—15 Multipara—26	35.5	2365	Yes	a	41	a	26 (63%)	19 (73%)	Yes	7	Fetal distress—2 Cord prolapse—3 Compound presentation—1 Failure to descend—1	15 (37%)	CC—9 Breech extraction—6
3.	Wells et al. 1991 23	25.4	Nullipara—9 Multipara—14	35.3	2389	a	No	23	a	11 (48%)	11 (100%)	a	0	‐	12 (52%)	CC—11 (including: unsuccessful ECVs—8, cord prolapse during ECV—3, fetal distress—1) Breech extraction—1
4.	Adam et al. 1991 13	a	a	a	a	a	a	15	a	12 (80%)	12 (100%)	a	0	‐	3 (20%)	Breech extraction—2 Breech delivery—1
5.	Greig et al. 1992 19	a	a	a	a	a	a	16	a	10 (63%)	10 (100%)	a	0	‐	6 (37%)	CC—3 Breech extraction—3
6.	Chauchan et al. 1995 22	24.3	a	33.0	2093	Yes	Yes	21	Breech—13 Transverse—7 Oblique—1	10 (48%)	10 (100%)	Yes	0	‐	11 (52%)	CC—10 (including: unsuccessful ECVs—6, cord prolapse—2, abnormal FHR—2) Breech extraction—1
7.	Kaplan et al. 1994 18	a	a	N/A	N/A	Yes	No	96	a	72 (75%)	72 (100%)	Yes	0	‐	24 (25%)	CC—2 Breech extraction—22
8.	Smith et al. 1997 16	30.6	Nullipara—14 Multipara—19	36.1	2495 g	Yes	Yes	33	Breech—17 Transverse—14 Oblique—2	17 (52%)	13 (77%)	Yes	4	Cord prolapse—2 Fetal distress—2	16 (48%)	CC—4 Breech extraction—12
9.	Mauldin et al. 1998 17	24.4	a	35.0	2295 g	a	a	19	a	8 (42%)	6 (75%)	a	2	Fetal distress—1 Abruptio placenta—1	11 (58%)	CC—10 (including: fetal distress—3, unsuccessful ECVs—7) Breech extraction—1

^a Not reported.

3.3. Adverse maternal and neonatal outcomes

We decided to classify neonatal death, major birth trauma, and Apgar score <7 at 5 min as adverse neonatal outcomes. Seven out of ten authors covered this matter in their studies. ¹⁵ , ¹⁶ , ¹⁷ , ²⁰ , ²¹ , ²² , ²³ In one case, information on birth trauma was incomplete. ¹⁵

Overall, no cases of major birth trauma and only one case of fetal death were reported. ²¹ That was a preterm neonate (28 weeks of gestation), in whom successful ECV was performed, resulting in a vertex vaginal delivery. The death occurred 14 days postpartum, due to necrotizing enterocolitis. Regarding neonatal outcomes, 8 out of 192 neonates (4.2%) had the Apgar score <7 at 5 min after birth.

Adverse maternal outcomes were reported in 6 studies. ¹⁵ , ¹⁶ , ¹⁷ , ²⁰ , ²¹ , ²³ Ultimately, one study was excluded from further analysis because the results were presented as the composite outcome and the absolute number of any specific complications could not be determined. ¹⁷

The summary of the results of adverse neonatal and maternal outcomes is shown in Table 2.

TABLE 2.

Adverse neonatal and maternal outcomes.

	Authors and year	Number of ECVs	Maternal adverse outcomes	5 min Apgar <7	Adverse neonatal outcomes
1.	Chervenak at el. 1983 21	25	Uterine atony—1 Endometritis—2	2	Neonatal death—1 (NEC) Birth trauma—0
2.	Gocke et al. 1989 20	41	PPH—1 Endometritis—2	0	Neonatal death—0 Birth trauma—0
3.	Wells et al. 1991 23	23	Transfusion—2	1	Neonatal death—0 Birth trauma—0
4.	Tchabo et al. 1992 15	30	PPH—1 Uterine inversion—1 Endometritis—1	0	Neonatal death—0 Birth trauma—not reported
5.	Chauchan et al. 1995 22	21	Not reported	0	Neonatal death—0 Birth trauma—0
6.	Smith et al. 1997 16	33	Endometritis—4	3	Neonatal death—0 Birth trauma—0
7.	Mauldin et al. 1998 17	19	Cannot determine	2	Neonatal death—0 Birth trauma—0

Abbreviations: NEC, necrotizing enterocolitis; PPH, postpartum hemorrhage.

4. DISCUSSION

In our series, the overall success rate of ECV was 64.4%. Similar results were reported in studies conducted in the population of singletons. In one of the largest meta‐analyses covering 12.955 procedures, the pooled success rate was 58%. ¹¹ Clinical research on ECV provided valuable information, showing that higher success rates were associated with the presence of certain features or the application of specific interventions. Multigravidity, higher levels of amniotic fluid index, a complete breech presentation, reduced uterine tension, or tocolysis were the most commonly reported predictive factors. ²⁴ , ²⁵ , ²⁶ , ²⁷ , ²⁸ , ²⁹ , ³⁰ Regrettably, scant data confronting this matter are available for the population of twins, and only few authors in our series actually addressed this aspect. Even though they demonstrated no association between success rates and parity, gestational age, or neonatal birthweight, none of them performed any deeper analysis. ¹⁶ , ¹⁹ , ²¹ Similarly, some of them reported interventions such as the administration of epidural anesthesia or tocolysis at the time of ECV, but no clear conclusions were made regarding their influence on the overall success rates. ¹⁶ , ¹⁸ , ²⁰ , ²¹ , ²² , ²³

In a subgroup of successful versions, cephalic vaginal delivery was achieved in 91.9% of cases. Interestingly, five studies showed that all successful versions were followed by vertex vaginal deliveries in their cohorts. ¹³ , ¹⁸ , ¹⁹ , ²² , ²³ Even though four cohorts were relatively small, the numbers in the remaining one were particularly high (72 successful ECVs, each followed by vertex vaginal delivery). ¹⁸ Compared to all other studies in our series, those outcomes seemed exceptionally surprising. Regrettably, information regarding factors that might influence the overall results was not provided.

Data regarding other modes of delivery among successful versions as well as unsuccessful ones (breech extractions and cesarean sections) were also limited, with no clear definitions of particular indications being provided for any form of management.

Although the majority of authors reported a relatively low incidence of adverse neonatal and maternal outcomes, due to the small sample size and significant differences in particular definitions, no reliable conclusions can be made.

To our best knowledge, this is the first systematic review that addresses the application of ECV in this particular population and investigates the events following the procedure. We strongly believe that ECV should not be considered as a stand‐alone procedure but rather as an integral part of the overall management. Notably, our main goal is to achieve vertex vaginal delivery, not merely a simple change in the fetal presentation. In our opinion, evaluating the entire sequence of events would provide more insight into the actual utility of this procedure.

A low number of studies that actually addressed this issue was one of the key problems that we encountered during the literature search for this review. In the final step, we included 10 studies, with the latest manuscript being published in 1998. This translates not only into the lack of data concerning the past 20 years, but also due to the retrospective nature and poor methodology of research at that time, it makes the reported results hard to compare. ¹³ , ¹⁵ , ¹⁶ , ¹⁷ , ¹⁸ , ¹⁹ , ²⁰ , ²¹ , ²² , ²³ For example, in cases of neonatal or maternal outcomes, the reported results always referred to the entire group of pregnancies in which ECV was performed. We were unable to categorize those results (neonatal outcomes for pregnancies with successful vs. unsuccessful ECVs), so no precise conclusions could be made. Moreover, the definitions of maternal outcomes varied significantly between studies, not allowing their comparison.

The existing knowledge on the ECV procedure is predominantly obtained from studies conducted in the population of singletons, which is largely a natural consequence of the low prevalence of twin pregnancies. However, another possible factor, particularly associated with the low number of ECVs in twins, could be related to the fact that, regardless of fetal configuration, the absolute risk of complications is higher in multiple pregnancies than in singletons. ³¹ This observation, as well as reports from numerous retrospective studies suggesting that planned cesarean delivery reduced those risks, encouraged numerous clinicians to abandon trials of vaginal delivery in twins. Therefore, possible ECVs were also abandoned, and cesarean sections were opted for regardless of the fetal presentation. ⁶ , ⁷ , ⁸

Clinical data regarding ECV performed on the second twin are significantly less available than the amount of research conducted in singletons. The general aspects, such as the success rate, mode of delivery, and adverse outcomes, were largely covered, with the results being promising. However, poor methodology and the retrospective nature of those studies made it difficult to draw reliable conclusions for clinical management. Another extremely important concern is related to the lack of data covering the past 20 years. Although the main findings, especially rates of successful procedures and adverse outcomes, seem to be similar to the results obtained in singletons, it must be emphasized that the direct comparison of both groups is impossible. The main reason stems from the different nature of the procedure performed in those groups. In singletons, ECV is primarily planned, whereas in nonvertex second twins, it is performed intrapartum, after the delivery of the first fetus. Furthermore, in the majority of cases, fetal presentation in singletons at term is relatively stable, with breech being the most common type in the nonvertex subgroup. ³² Considering this matter in twin pregnancies, the situation is much more complex because the presence and delivery of the first fetus may influence the final presentation of the second twin, which could be different from the initial one. ¹³ Taking account of the fact that multifetal pregnancies are at a higher risk of overall pregnancy complications compared to singletons, those different conditions during ECV could significantly bias the final results. ³¹

Finally, it is important to mention that apart from ECV, other ways of the management of nonvertex second twins are also available, including spontaneous breech delivery, internal podalic version, and breech extraction. Future research should focus on the deeper analysis of such methods and their comparison with ECV.

5. CONCLUSION

The management of vaginal delivery in twins is still challenging, especially in case of the nonvertex presentation of the second fetus. The results of our review suggest that the intrapartum ECV of nonvertex second twins might be a reasonable option for clinical management. However, due to the scarcity and low quality of available data, it is still impossible to provide any general conclusions. Further research on this topic, with particular attention paid to randomized trials comparing other management methods, is needed to provide detailed information. This will enable patients to make better informed choices, enhancing their knowledge on this topic, and help clinicians formulate clear recommendations for future clinical management.

AUTHOR CONTRIBUTIONS

Miłosz Dymon: Design, extraction and synthesis of the data, writing the first draft. Marta Szajnik: Design. Małgorzata Siergiej: Literature search. Jakub Kociuba: Literature search. Magdalena Zgliczyńska: Extraction and synthesis of the data, critical revisions of successive drafts. Michał Ciebiera: Critical revisions of successive drafts. All authors agreed to publish the paper.

CONFLICT OF INTEREST STATEMENT

The authors have no conflict of interest to declare.

Supporting information

Appendix S1. Search strategy.

Figure S2. The PRISMA flow diagram.

Table S3. Main characteristics of included studies.

Table S4. Risk of bias.

Dymon M, Ciebiera M, Zgliczyńska M, Siergiej M, Kociuba J, Szajnik M. External cephalic version in nonvertex second twin—Success rate, mode of delivery, and safety: A systematic review. Acta Obstet Gynecol Scand. 2025;104:584‐590. doi: 10.1111/aogs.15006