The rate of twin pregnancies continues to rise, from an incidence of approximately 20 per 1,000 live births in the 1990s to more than 30 per 1,000 live births in 2009 in the United States and Canada.1,2 1,2 Recent evidence supports vaginal delivery as the preferred mode of delivery when twin A is vertex, irrespective of the presentation of twin B.3–6 3–6 3–6 3–6 However, some care providers and departmental policies in certain centers may support cesarean delivery as the preferred mode of delivery when twin B is in the nonvertex presentation as a result of either lack of adequate experience in maneuvers such as total breech extraction and internal podalic version or medicolegal considerations.
In singleton pregnancies, fetal presentation is unlikely to change beyond 36 weeks of gestation.7 However, data regarding the likelihood of spontaneous change in fetal presentation during the third trimester in twin pregnancies are limited because this question has been addressed by only a small number of studies.8–11 8–11 8–11 8–11 In addition, available studies are limited by a relatively small sample size and variation in the gestational age at the time of ultrasound assessment and provide very little information regarding factors that are predictive of spontaneous change in the presentation of either of the twins.
Our aim was to estimate the likelihood of spontaneous change in the presentation of either twin during the third trimester and to identify predictors of spontaneous fetal version in a large cohort of twin pregnancies.
MATERIALS AND METHODS
This study was based on a secondary analysis of the data from a multicenter randomized controlled trial on mode of delivery in twin pregnancies.4 Women were enrolled in the original study if they were between 32 and 38 weeks of gestation, the first twin was in the cephalic presentation, and both twins were alive with an estimated weight between 1,500 g and 4,000 g confirmed by means of ultrasonography within 7 days of randomization. Exclusion criteria were monoamniotic twins, fetal reduction at 13 or more weeks of gestation, the presence of a lethal fetal anomaly, contraindication to labor or vaginal birth (eg, fetal compromise, second twin substantially larger than the first twin, fetal anomaly or condition that might cause mechanical problems at delivery, and previous vertical uterine incision or more than one previous low-segment cesarean delivery), and previous participation in the study. Participants were randomly assigned to planned cesarean delivery or planned vaginal birth. Data were abstracted from the medical records at participating centers by trained study staff and were recorded, after delivery, on standardized data collection forms. Elective delivery by means of either cesarean delivery (for women in the planned cesarean delivery group) or labor induction (for women in the planned vaginal birth group) was planned between 37 weeks 5 days and 38 weeks 6 days of gestation. Overall the study included 2,804 women (1,398 in the planned cesarean delivery group and 1,406 in the planned vaginal delivery group) who were recruited from 106 centers in 25 countries. The original study and all secondary analyses were approved by the research ethics board at the Sunnybrook Health Sciences Centre.
Data were abstracted from the study database and included the following information: maternal demographics and obstetric history, chorionicity, gestational age at the time of randomization, presentation and estimated weight of both twins at the time of ultrasound examination, gestational age at delivery, presentation of both twins at the time of delivery, birth weight, and fetal sex. Twin A was defined as the fetus closest to the maternal cervix. Breech or transverse presentation was documented as nonvertex.
In the current study we analyzed the likelihood of a spontaneous version of either of the twins during the third trimester between the ultrasound assessment at the time of randomization and delivery. The likelihood of change in the presentation of twin A and twin B was calculated and was stratified by the presentations at the time of ultrasound examination, the gestational age at the time of ultrasound examination, and by the ultrasound examination-to-delivery interval. The characteristics of fetuses that underwent spontaneous version were compared with those that remained in the same presentation. Multivariable logistic regression analysis was used to identify factors associated with spontaneous version of either of the twins while adjusting for potential confounding variables. The χ2 test and the Student's t test were used to compare categorical and continuous variables between the groups, respectively. Adjusted odds ratios plus 95% confidence intervals (CIs) were determined. All P values in all tables were two-sided. The level for declaring statistical significance was set to .05. Because the analysis was secondary to a randomized clinical trial, no power calculations were performed. The power of the analysis and corresponding results are reflected in the width of the CIs. Data analysis was performed with SPSS 21. This research adhered to the STrengthening the Reporting of OBservational studies in Epidemiology (STROBE) guidelines for observational studies.
A total of 2,603 women with twin pregnancies were included in the analysis (Table 1). The presentation of the twins at the time of randomization was vertex–vertex in the majority of cases (56.9%, 95% CI 55.0–58.8%) followed by vertex–breech (26.3%, 95% CI 24.6–28.0%) and vertex–transverse (16.9%, 95% CI 15.4–18.4%).
The overall likelihood of change of presentation of the twins is presented in Figure 1. The vertex–vertex presentation was the most stable combination and persisted at the time of delivery in 81.4% of cases compared with 70% in the case of vertex–breech presentation and 55.8% in the case of vertex–transverse presentation, which was the least stable combination.
Twin A tended to persist in the vertex presentation after 32 weeks of gestation with an overall rate of spontaneous version to nonvertex presentation of only 3.0% (Table 2). This rate was significantly lower in cases of vertex–vertex presentation (1.6%) than when twin B was in the breech presentation (5.4%, P<.001) or transverse lie (4.3%, P=.001) (Fig. 1).
Twin B was much less stable and underwent spontaneous version in 24.8% of cases (Table 2). The rate was significantly lower when twin B was in the vertex presentation (17.1%) than when it was in breech presentation (24.6%, P<.001) or transverse lie (39.9%, P<.001) at the time of randomization (Fig. 1).
Figure 2 presents the likelihood of spontaneous version based on the gestational age at the time of randomization and the randomization-to-delivery interval. Twin B was more likely to change its presentation as the randomization-to-delivery interval increased (Fig. 2A). The likelihood of spontaneous version of twin B decreased with more advanced gestational age at the time of randomization, although the rate remained higher than 20% even after 34 weeks of gestation (Fig. 2B). This relationship was less pronounced in the case of twin A—the likelihood of spontaneous version of twin A to a nonvertex presentation increased only when the randomization-to-delivery interval was greater than 5 weeks (Fig. 2A).
We next compared the characteristics of pregnancies in which twin A has changed his or her presentation between the time of randomization and delivery compared with those pregnancies in which twin A remained in the vertex presentation (Table 2). In those pregnancies in which twin A changed its presentation, twin B was more likely to be in a nonvertex presentation (breech presentation or transverse lie) and to have a lower estimated weight at the time of randomization, and the randomization-to-delivery interval was greater compared with pregnancies in which twin A remained vertex (Table 2).
With respect to twin B, pregnancies in which twin B changed its presentation from the time of randomization to the time of delivery were characterized by a higher maternal age, lower gestational age at randomization, greater randomization-to-delivery interval and a higher gestational age at delivery, and a lower estimated weight of both twin A and twin B at the time of randomization compared with pregnancies in which twin B did not change its presentation (Table 2). In addition, pregnancies in which twin B underwent spontaneous version were more likely to involve multiparous women, dichorionic–diamniotic twins, and a nonvertex presentation of twin B at the time of randomization (Table 2).
Multivariable logistic regression analysis was used to identify factors that are independently associated with change in presentation of either of the twins (Table 3). Factors that were found to be associated with a spontaneous version of twin A to nonvertex presentation were twin B being in the nonvertex presentation, lower weight estimation of twin B, change in presentation of twin B before delivery, and an interval of more than 4 weeks between randomization and delivery (Table 3).
Factors that were associated with a change in the presentation of twin B varied with the original presentation of twin B. When twin B was in the vertex presentation, factors that were associated with a spontaneous version to nonvertex presentation included a randomization-to-delivery interval of more than 4 weeks and multiparity (Table 3). Factors associated with a spontaneous version of twin B from the breech presentation or transverse lie to vertex presentation were randomization-to-delivery interval of more than 4 weeks and a spontaneous version of twin A (Table 3).
To present the information in a manner that is more relevant for clinical practice that can be used for the purpose of counseling, we calculated the likelihood of change in twin presentations based on the initial combination of twin presentations and the predictive factors presented (see Appendix 1, available online at http://links.lww.com/AOG/A714).
The aim of the current study was to determine the likelihood of spontaneous change in the presentation of twins during the third trimester in a large cohort of twin pregnancies. Our main findings are as follows: 1) twin A tends to remain stable in the vertex presentation after 32 weeks of gestation with a likelihood of spontaneous version to nonvertex presentation of only approximately 3%; 2) the presentation of twin B is less stable and changes in approximately 25% of cases—this remains true even when twin B is in the vertex presentation (version rate of 17%) and even late in the third trimester (version rate of higher than 20% after 34 weeks of gestation); 3) both twins are more likely to change their presentation when twin B is in the nonvertex presentation and is smaller, and as the interval to delivery after the ultrasound examination increases; 4) twin A is also more likely to switch to nonvertex presentation when twin B changes its presentation; and 5) the presentation of twin B is less stable in multiparous women.
Information regarding the likelihood of spontaneous version of twins during the third trimester has been addressed previously by only a small number of studies,8–11 8–11 8–11 8–11 which were limited by relatively small sample sizes (n=119–491) and by limiting the analysis only to the presenting twin.10
We have found that twin A tends to remain stable after 32 weeks of gestation and that the vertex–vertex presentation is the most stable combination. These findings are in agreement with previous studies. Fox et al,11 in a study of 441 twin pregnancies, found that when twin A is in the vertex presentation at 28–32 or 32–36 weeks of gestation, this presentation will persist to term in 90–95% and 96–97% of cases, respectively. In a similar study of 238 twin pregnancies, Schwartz et al10 found that a vertex presentation of twin A after 28 weeks of gestation persisted to term in 95% of cases. Finally, Chasen et al9 reported that a vertex presentation of twin A persisted in 95% and 98% of cases when a vertex presentation was documented at greater than 28 weeks and greater than 32 weeks of gestation, respectively. In agreement with our findings, Divon et al8 reported that the vertex–vertex presentation was the most stable combination that changed in only 7% of cases.
Data regarding the likelihood of change in the presentation of twin B are more limited. We have found that the rate of spontaneous version of twin B is high (24.8%), even when twin B is in the vertex presentation and even late during the third trimester. Similar to our findings, the rate of change in the presentation of twin B after 32 weeks of gestation was 32% in the study of Fox et al11 and 27% in the study of Chasen et al.9
Data regarding the factors that affect the likelihood of version of twin fetuses are even more limited. We identified several factors that are independently associated with spontaneous version of twin A, including a nonvertex presentation of twin B, lower weight of twin B, change in presentation of twin B, and an interval to delivery of more than 4 weeks. In contrast to our findings, Fox et al11 found that the factors that were associated with a spontaneous version of twin A were multiparity and a higher estimated weight percentile of either twin A or twin B. One possible reason for these conflicting results may be the fact that in this latter study, the authors included all cases of version of twin A as the outcome variable (ie, either vertex to breech or breech to vertex with the majority of cases being breech to vertex version), whereas in the current study, the analysis was limited to only cases of version from vertex to breech presentation. Thus, it is possible that the factors that predispose to change in the presentation of twin A vary based on the initial fetal presentation so that larger twins are more likely to be forced into the vertex presentation, whereas a smaller twin B may allow more space for twin A to switch from the vertex to the breech position. Another possible explanation may be the lack of adjustment for potential confounders (eg, gestational age at ultrasound examination) in the study of Fox et al.11 Two other studies were not able to identify any predictors for spontaneous version, including parity, fetal weight, chorionicity, fetal sex, placental location, or amniotic fluid level, although this may be attributed to the lower power of these studies.8,9 8,9
None of the available studies identified factors that are predictive of version of twin B. In the current study we found that the factors that are associated with version of twin B are similar to those identified for twin A (ie, twin B being smaller and in the nonvertex presentation and an interval to delivery of more than 4 weeks). In addition, we found that some predictors varied with the original presentation of twin B. Thus, twin B was more likely to switch from vertex to nonvertex presentation in multiparous women, possibly as a result of the greater relaxation of the abdominal wall and pelvic floor muscles. In contrast, twin B was more likely to switch from nonvertex to vertex presentation after spontaneous version of twin A to the nonvertex presentation.
The main limitation of our study is that the study population was limited to pregnancies with twin A in vertex presentation; thus, we were not able to determine the likelihood of a change of spontaneous version of twin A from nonvertex to vertex presentation. In addition, information on the presentation of the twins at the time of delivery was taken from the delivery report and was not necessarily confirmed by ultrasonography immediately before delivery.
In summary, we have found that twin A, when in the vertex presentation, is unlikely to change its presentation after 32 weeks of gestation, but that twin B can change its presentation in approximately 25% of cases, even late in pregnancy.
1. Fell DB, Joseph K. Temporal trends in the frequency of twins and higher-order multiple births in Canada and the United States. BMC Pregnancy Childbirth 2012;12:103.
2. Martin JA, Hamilton BE, Osterman MJ. Three decades of twin births in the United States, 1980–2009. NCHS Data Brief 2012:1–8.
3. Fox NS, Silverstein M, Bender S, Klauser CK, Saltzman DH, Rebarber A. Active second-stage management in twin pregnancies undergoing planned vaginal delivery in a U.S. population. Obstet Gynecol 2010;115:229–33.
4. Barrett JF, Hannah ME, Hutton EK, Willan AR, Allen AC, Armson BA, et al.. A randomized trial of planned cesarean or vaginal delivery for twin pregnancy. N Engl J Med 2013;369:1295–305.
5. Mauldin JG, Newman RB, Mauldin PD. Cost-effective delivery management of the vertex and nonvertex twin gestation. Am J Obstet Gynecol 1998;179:864–9.
6. Schmitz T, Carnavalet Cde C, Azria E, Lopez E, Cabrol D, Goffinet F. Neonatal outcomes of twin pregnancy according to the planned mode of delivery. Obstet Gynecol 2008;111:695–703.
7. Hill LM. Prevalence of breech presentation by gestational age. Am J Perinatol 1990;7:92–3.
8. Divon MY, Marin MJ, Pollack RN, Katz NT, Henderson C, Aboulafia Y, et al.. Twin gestation: fetal presentation as a function of gestational age. Am J Obstet Gynecol 1993;168:1500–2.
9. Chasen ST, Spiro SJ, Kalish RB, Chervenak FA. Changes in fetal presentation in twin pregnancies. J Matern Fetal Neonatal Med 2005;17:45–8.
10. Schwartz R, Fuchs A, Rosenn B. Can third trimester ultrasound predict the presentation of the first twin at delivery? J Matern Fetal Neonatal Med 2012;25:2432–4.
11. Fox NS, Rebarber A, Lesser HN, Roman AS, Klauser CK, Saltzman DH. Factors affecting fetal presentation in twin pregnancies across gestation. J Matern Fetal Neonatal Med 2013;26:1658–61.