The effects of labor and delivery on maternal and neonatal outcomes in term twins: a retrospective cohort study

Authors

  • DJ Wenckus,

    Corresponding author
    1. Division of Maternal Fetal Medicine, Department of Obstetrics and Gynecology, University of Illinois at Chicago, Chicago, IL, USA
    • Correspondence: Dalia J. Wenckus, MD, Wright Patterson Medical Center, Department of Obstetrics and Gynecology, Maternal Fetal Medicine, 4881 Sugar Maple Drive, Wright Patterson Air Force Base, OH 45433, USA. Email: dalia.wenckus@mac.com

    Search for more papers by this author
  • W Gao,

    1. Center for Clinical and Translational Science, University of Illinois at Chicago, Chicago, IL, USA
    Search for more papers by this author
  • MA Kominiarek,

    1. Division of Maternal Fetal Medicine, Department of Obstetrics and Gynecology, University of Illinois at Chicago, Chicago, IL, USA
    Search for more papers by this author
  • I Wilkins

    1. Division of Maternal Fetal Medicine, Department of Obstetrics and Gynecology, University of Illinois at Chicago, Chicago, IL, USA
    Search for more papers by this author

Abstract

Objective

To compare maternal and neonatal outcomes in twins undergoing a trial of labor versus pre-labor caesarean.

Design

Retrospective cohort study.

Setting

19 US hospitals from the Consortium on Safe Labor.

Population

Of 2225 twin sets ≥36 weeks' gestation.

Methods

Maternal (abruption, estimated blood loss, postpartum haemorrhage, transfusion, chorioamnionitis, hysterectomy, ICU admission, death) and neonatal outcomes (birth injury, 5-minute Apgar <7, NICU admission, RDS, TTN, sepsis, asphyxia, NICU length of stay, death) were compared between the trial of labour and pre-labour caesarean groups with univariate and multivariate logistic and linear regression analyses. Similar analyses were performed for actual delivery modes.

Main outcome measures

Maternal and neonatal outcomes.

Results

Among the 2225 twin sets, 1078 had a trial of labour, and 65.9% of those delivered vaginally. There was an increased risk for postpartum haemorrhage [OR 2.5, 95% confidence interval (CI) 1.4–4.5] and blood transfusion (OR 1.9, 95%CI 1.2–3.2) for the trial of labour compared with pre-labour caesarean groups. Birth injury only occurred in the trial of labour group, 1% Twin A, 0.4% Twin B. Both twins had a higher risk of 5-minute Apgar <7 with trial of labour compared to pre-labour caesarean (A: OR 3.9, 95%CI 1.05–14.5; B: OR 3.9, 95%CI 1.3–12.3).

Conclusion

Term twins undergoing a trial of labour have increased maternal haemorrhage and transfusions along with neonatal birth trauma and lower Apgar scores, but these absolute neonatal occurrences were rare. Trial of labour in twins remains a safe and reasonable option in appropriately selected cases.

Introduction

The optimal delivery mode for twin gestations remains controversial. The American College of Obstetricians and Gynecologists (ACOG) practice bulletin on multiple gestations, states that fetal position, ability to monitor fetal heart rate, and maternal and fetal status should determine delivery mode for twins.[1] In 1989, 50% of twin gestations were delivered by caesarean, and this number had increased to 75% by 2008.[2, 3] This high caesarean delivery rate for twin gestations begs the question of whether a trial of labour for twins has unacceptable risks for the mother and neonate.

In singletons, pre-labour and elective repeat caesarean deliveries increase the risk for several adverse maternal and neonatal outcomes over vaginal delivery despite the controlled circumstances in which they are performed, and these risks increase with each subsequent caesarean delivery.[4-6] A trial of labour is generally the safest option for singletons, but increased maternal and neonatal morbidity and mortality exist when caesarean delivery occurs after a trial of labour or when conducted in an emergent fashion.[7-9] It is unknown whether the risks and benefits of a trial of labour also apply to twin gestations.

Previous investigations found no significant differences in maternal and first twin outcomes when comparing planned trial of labour to planned caesarean delivery, though few studies directly addressed maternal outcomes.[10-12] In light of the increased morbidity and mortality in the second twin after a planned trial of labour in larger population studies, some investigators have recommended a planned caesarean delivery.[13-16] Two recent reports found no significant differences in neonatal outcomes for the second twin according to planned delivery mode. As such, they recommended a trial of labour for twins with the goal of vaginal delivery.[17, 18]

We compared maternal and neonatal outcomes for twin gestations delivered after a trial of labour with those who had a pre-labour caesarean. Our hypothesis was that increased maternal and neonatal morbidity exists in the trial of labour group. Additionally, we compared actual delivery mode after trial of labour with pre-labour caesarean to search for the presence of an outcome so catastrophic that it would preclude offering a particular delivery mode in twins. Our secondary hypothesis was that maternal and neonatal morbidity would not increase for vaginal deliveries and any additional morbidity would be limited to caesareans after labour.

Methods

This is a retrospective cohort study from the Consortium on Safe Labor (CSL). The complete database contained 233 730 deliveries from 2002 to 2008, acquired from electronic obstetrical databases. Twelve clinical centres from 19 distinct hospitals across nine ACOG districts participated in the CSL. Detailed description of the study is provided elsewhere.[19, 20] Institutional Review Boards of all participating institutions approved the study.

The entire CSL database included 4864 twin sets, encompassing 2.1% of all deliveries. For the current cohort study, the inclusion criteria were twins that were both liveborn and ≥36 weeks' gestation, for a total of 2225 twin sets. Antepartum stillbirths were excluded because the aim of the study was to measure the effects of labour. There were no intrapartum stillbirths. Comparison groups were determined based on their exposure to labour. A trial of labour (TOL) was defined as a vaginal delivery or at least two cervical examinations documented in the obstetrical database during the labour admission. The TOL group was further separated into vaginal deliveries (TOL:vaginal) and caesarean deliveries (TOL:caesarean). The pre-labour caesarean group (PLC) was defined as a caesarean delivery and ≤1 cervical examination documented.

Characteristics of the cohort recorded included age, race, parity, prior caesarean, admission body mass index (BMI), diabetes, hypertensive disorders, gestational age at delivery, and birthweight. Maternal outcomes included abruption, estimated blood loss (EBL), postpartum haemorrhage, blood transfusion, chorioamnionitis, hysterectomy, ICU admission, and death prior to discharge. Neonatal outcomes included birth injury (brachial plexus injury, skull fracture, facial nerve injury, laceration, clavicular fracture, shoulder dislocation), 5-minute Apgar <7, neonatal intensive care unit (NICU) admission, respiratory distress syndrome (RDS), transient tachypnea of the newborn (TTN), neonatal sepsis, asphyxia, NICU length of stay, and death prior to discharge.

Maternal and neonatal characteristics were compared with chi-square and Student's t-tests for categorical and continuous variables, respectively for the TOL versus PLC groups. Maternal and neonatal outcomes were compared for the TOL versus PLC groups. These outcomes were also compared between the TOL:vaginal versus PLC groups, and the TOL:caesarean versus PLC groups to further delineate where morbidity occurs. Combined vaginal-caesarean deliveries remained within the TOL:vaginal delivery group for analysis because the intent was to deliver vaginally. Median and interquartile ranges (IQR) were reported for continuous outcome variables as none was normally distributed. The medians were compared with the Wilcoxon rank sum test. Univariate analysis with simple logistic regression calculated unadjusted odds ratios with 95% confidence intervals (CI) for each of the maternal and neonatal outcomes. Multivariate analyses were conducted with logistic regression calculating adjusted odds ratios with 95%CI, after controlling for age, parity, admission BMI, prior caesarean, and diabetes. A P-value <0.05 was considered statistically significant for all comparisons. sas software (version 9.2; SAS Institute Inc., Cary, NC, USA) was used for all analyses.

Results

Of the 2225 twins studied, 1078 had a TOL. In the TOL group, 710 (65.9%) had a vaginal delivery, 311 (28.9%) a caesarean delivery, and 57 (5.3%) a combined vaginal-caesarean delivery. There were 1147 women (51.6% of the cohort) who had a PLC. Overall, caesarean deliveries occurred in 68.1% of this twin study cohort.

The PLC group was more likely to be older, have a prior cesarean, diabetes, and a higher admission BMI, < 0.05 (Table 1). The mean gestational age at delivery was 37.3 weeks for both groups. The most common indication for caesarean was ‘failure to progress' in the TOL group and ‘breech' in the PLC group. It could not be determined to which birth order fetus the term ‘breech' belonged.

Table 1. Maternal characteristics
Variable (SD or %)Trial of labour (TOL) n = 1078Pre-labour caesarean (PLC) n = 1147P-value
Mean age (years) 29.4 (±6.1)31.4 (±6.6)<0.001
Race
White610 (59.0)663 (60.3)0.11
African American234 (22.6)210 (19.1)
Hispanic143 (13.8)158 (14.4)
Other47 (4.6)68 (6.2)
Nullipara 426 (39.5)497 (43.3)0.07
Prior caesarean 54 (5.5)289 (25.9)<0.001
Diabetes 60 (5.6)96 (8.4)0.01
Hypertensive disorders 150 (13.9)135 (11.8)0.13
Mean admission BMI (kg/m 2 ) 32.3 (±6.7)33.4 (±7.1)0.001
Caesarean delivery 311 (28.9)  
Mean gestational age (weeks) 37.3 (±1.0)37.3 (±0.9)0.86
Birth weight (g)
Twin A2687 (±358)2716 (±387)0.07
Twin B2659 (±392)2661 (±405)0.89

Table 2 details the univariate and multivariate analyses of maternal and neonatal outcomes for the TOL group compared with the PLC group. The median EBL for the TOL group was 500 ml (IQR 400–800 ml) and the median EBL for the PLC group was 800 ml (IQR 700–1000 ml), < 0.001. The TOL group required more blood transfusions than the PLC group (aOR 1.9, 95%CI 1.2–3.2). Postpartum haemorrhage increased for TOL (aOR 2.5, 95%CI 1.4–4.5) compared with PLC. There were no differences in ICU admissions and postpartum hysterectomies for TOL versus PLC. One maternal death occurred in the TOL group. She was a 26-year-old gravida 1 with a history of depression and pre-eclampsia admitted at 37 weeks for delivery, given a beta-blocker for blood pressure management and delivered via vacuum-assisted vaginal delivery of Twin A and Twin B, with an estimated blood loss of 500 ml. No further information is available in the database on this event.

Table 2. Maternal and neonatal outcomes for TOL vs PLC
 TOL n = 1078 (%)PLC n = 1147 (%)Unadjusted OR 95% CI)Adjusted OR (95% CI)
  1. a

    Regression model does not fit, due to low outcome occurrence.

Maternal outcomes
Abruptio Placentae10 (1.2)1 (0.1)10.9 (1.4–85.1)8.6 (1.1–69.7)
Estimated blood loss ≥1000 ml120 (14.7)264 (29.1)0.4 (0.3–0.5)0.5 (0.4–0.7)
Postpartum 1 haemorrhage53 (6.6)20 (2.7)2.5 (1.5–4.3)2.5 (1.4–4.5)
Blood transfusion55 (5.7)38 (3.7)1.6 (1.02–2.4)1.9 (1.2–3.2)
Chorioamnionitis38 (4.5)2 (0.2)21.3 (5.1–88.8)23.1 (5.3–99.7)
Hysterectomy1 (0.1)2 (0.2)0.5 (0.05–5.9) a
ICU admission1 (0.1)6 (0.6)0.2 (0.02–1.5)0.3 (0.03–2.4)
Maternal death1 (0.1)0 a a
Neonatal outcomes
Birth injury
Twin A11 (1.0)0 a a
Twin B4 (0.4)0 a a
5 minute APGAR <7
Twin A13 (1.2)7 (0.6)2.0 (0.8–5.0)3.9 (1.05–14.5)
Twin B17 (1.6)7 (0.6)2.6 (1.1–6.4)3.9 (1.3–12.3)
NICU admission
Twin A154 (14.3)164 (14.3)1.0 (0.8–1.3)1.1 (0.8–1.4)
Twin B170 (15.8)162 (14.1)1.1 (0.9–1.4)1.0 (0.8–1.4)
Respiratory distress syndrome
Twin A22 (2.0)23 (2.0)1.0 (0.6–1.8)1.0 (0.5–2.0)
Twin B31 (2.9)26 (2.3)1.3 (0.8–2.2)1.2 (0.6–2.2)
Transient Tachypnea of the newborn
Twin A34 (3.2)71 (6.2)0.5 (0.3–0.7)0.6 (0.3–0.9)
Twin B42 (3.9)67 (5.8)0.7 (0.5–0.97)0.5 (0.3–0.9)
Neonatal Sepsis
Twin A25 (2.3)16 (1.4)1.7 (0.9–3.2)1.8 (0.8–4.1)
Twin B19 (1.8)13 (1.1)1.6 (0.8–3.2)1.2 (0.5–2.8)
Neonatal Asphyxia
TwinA3 (0.3)1 (0.1)3.2 (0.3–30.8)5.0 (0.4–55.7)
Twin B2 (0.2)0 a a
NICU LOS  median
Twin A ≥4.0 days79 (59.0)99 (65.1)0.8 (0.5–1.2)1.0 (0.5–1.9)
Twin B ≥5.0 days76 (50.7)78 (53.1)0.9 (0.6–1.4)0.6 (0.3–1.1)
Neonatal death
Twin A2 (0.2)0 a a
Twin B2 (0.2)2 (0.2)1.0 (0.1–7.3)0.5 (0.1–3.8)

For neonatal outcomes (Table 2), birth injuries (brachial plexus injury, skull fracture, facial nerve injury, laceration, clavicular fracture, shoulder dislocation) only occurred in the TOL group, in 1% of Twin A and 0.4% of Twin B, and this was statistically significant for Twin A and B, P-values <0.001 and <0.05, respectively. Both twins had a higher risk of 5-minute Apgar <7 in the TOL group (Twin A aOR 3.9, 95%CI 1.05–14.5; Twin B aOR 3.9, 95%CI 1.3–12.3). TOL was protective for TTN (Twin A aOR 0.6, 95%CI 0.3–0.9; Twin B aOR 0.5, 95%CI 0.3–0.9). For twins admitted to the NICU, the median length of stay for Twin A in the TOL and the PLC groups was 4.0 days (IQR 2.0–8.0 days) and 4.3 days (IQR 3.0–8.5 days), respectively, = 0.53. For Twin B, the median length of stay for the TOL and PLC groups was 5.0 days (3.0–9.0 days) and 5.0 days (IQR 2.0–8.0), respectively, = 0.56.

Of the six cases of neonatal asphyxia, five occurred in the TOL group. The difference was not statistically significant compared to the PLC group, > 0.05. There were four neonatal deaths in the TOL group and two in the PLC group. The causes of death in the TOL group included GBS sepsis, congenital heart defect, hypoplastic lungs, and congenital anomaly. The causes of death in the PLC group (Twin B only) included perinatal asphyxia and congenital heart disease with sepsis.

Table 3 shows the univariate and multivariate analyses of maternal and neonatal outcomes for the TOL:vaginal group compared with the PLC group as well as analyses for the TOL:caesarean group compared with the PLC group. The median EBL was lower for the TOL:vaginal group (400 ml; IQR 300–600 ml) but similar for the TOL:caesarean group (800 ml; IQR 700–1000 ml), compared with the PLC group (800 ml; IQR 700–1000 ml), < 0.001 and = 0.92, respectively. Both TOL:vaginal (aOR 2.7, 95%CI 1.4–5.1) and TOL: caesarean (aOR 2.6, 95%CI 1.2–5.7) had increased postpartum haemorrhage compared with PLC; however, only TOL:caesarean had increased blood transfusions compared with PLC (aOR 3.2, 95%CI 1.8–5.9).

Table 3. Maternal and neonatal outcomes based on actual delivery mode compared to PLC
Maternal OutcomesTOL: vaginal n = 767 (%)Unadjusted OR (95% CI)Adjusted OR (95% CI)TOL: cesarean n = 311 (%)Unadjusted OR 95% CIAdjusted OR (95% CI)PLC n = 1147 (%)
Abruptio Placentae5 (0.8)7.6 (0.9–65.5)3.3 (0.4–31.9)5 (2.0)18.9 (2.2–162.4)17.6 (2.0–159.0)1 (0.1)
Estimated blood loss ≥1000 ml47 (8.1)0.2 (0.1–0.3)0.2 (0.1–0.4)73 (31.1)1.1 (0.8–1.5)1.2 (0.8–1.7)264 (29.1)
Postpartum Haemorrhage41 (6.8)2.6 (1.5–4.5)2.7 (1.4–5.1)12 (5.9)2.3 (1.1–4.8)2.6 (1.2–5.7)20 (2.7)
Blood transfusion32 (4.5)1.2 (0.8–2.0)1.3 (0.7–2.3)23 (8.7)2.5 (1.4–4.2)3.2 (1.8–5.9)38 (3.7)
Chorioamnionitis23 (3.8)18.1 (4.2–77.0)21.2 (4.6–98.2)15 (6.1)29.5 (6.7–130.0)22.6 (4.8–105.3)2 (0.2)
Hysterectomy0 b b 1 (0.3)1.8 (0.2–20.4) b 2 (0.2)
ICU admission1 (0.2)0.3 (0.03–2.1)0.4 (0.04–3.7)0 b b 6 (0.6)
Maternal death1 (0.2) b b 0 b b 0
Neonatal outcomesTOL: vaginal n = 767 sets (%)Unadjusted OR (95% CI)Adjusted OR (95% CI)TOL: cesarean n = 311 sets (%)Unadjusted OR (95% CI)Adjusted OR 95% CI)PLC n = 1147 sets (%)
  1. a

    P < 0.05.

  2. b

    Regression model does not fit, due to low outcome occurrence.

Birth injury
Twin A11 (1.4)a b b 0 b b 0
Twin B4 (0.5)a b b 0 b b 0
5 minute APGAR <7
Twin A12 (1.6)2.6 (1.02–6.6)6.4 (1.5–27.4)1 (0.3)0.5 (0.1–4.3)1.2 (0.1–11.2)7 (0.6)
Twin B16 (2.1)3.5 (1.4–8.5)6.1 (1.7–21.4)1 (0.3)0.5 (0.1–4.3)1.0 (0.1–8.9)7 (0.6)
NICU admission
Twin A97 (12.7)0.9 (0.7–1.1)0.9 (0.6–1.3)57 (18.3)1.3 (0.97–1.9)1.3 (0.9–2.0)164 (14.3)
Twin B115 (15.0)1.1 (0.8–1.4)1.0 (0.7–1.4)55 (17.7)1.3 (0.9–1.8)1.0 (0.7–1.6)162 (14.1)
Respiratory distress syndrome
Twin A18 (2.4)1.2 (0.6–2.2)1.2 (0.5–2.7)4 (1.3)0.6 (0.2–1.9)0.7 (0.2–2.3)23 (2.0)
Twin B25 (3.3)1.5 (0.8–2.5)1.5 (0.8–3.0)6 (1.9)0.8 (0.3–2.1)0.6 (0.2–1.8)26 (2.3)
Transient tachypnea of the newborn
Twin A18 (2.4)0.4 (0.2–0.6)0.4 (0.2–0.8)16 (5.1)0.8 (0.5–1.4)0.9 (0.5–1.6)71 (6.2)
Twin B21 (2.7)0.5 (0.3–0.7)0.3 (0.2–0.6)21 (6.8)1.2 (0.7–1.9)1.0 (0.5–1.8)67 (5.8)
Neonatal Sepsis
Twin A13 (1.7)1.2 (0.6–2.5)1.6 (0.6–4.1)12 (3.9)2.8 (1.3–6.1)2.2 (0.9–5.7)16 (1.4)
Twin B12 (1.6)1.4 (0.6–3.1)1.3 (0.5–3.2)7 (2.3)2.0 (0.8–5.1)0.9 (0.2–3.3)13 (1.1)
Neonatal Asphyxia
Twin A2 (0.3)3.0 (0.3–33.1)6.1 (0.3–107.9)1 (0.3)3.7 (0.2–59.3)5.2 (0.3–88.4)1 (0.1)
Twin B1 (0.1) b b 1 (0.3) b b 0
NICU LOSmedian
Twin A ≥4.0 days49 (57.0)0.7 (0.4–1.2)1.1 (0.5–2.3)30 (62.5)0.9 (0.5–1.7)0.8 (0.4–2.0)99 (65.1)
Twin B ≥5.0 days58 (57.4)1.2 (0.7–2.0)0.7 (0.4–1.5)18 (36.7)0.5 (0.3–1.0)0.3 (0.1–0.8)78 (53.1)
Neonatal death
Twin A2 (0.3) b b 0 b b 0
Twin B2 (0.3)1.4 (0.2–10.2)0.6 (0.1–4.4)0 b b 2 (0.2)

For neonatal outcomes, all of the birth injuries occurred in the TOL:vaginal group but, to reiterate, the occurrence was low – 1% for Twin A and 0.4% for Twin B. The 5-minute APGAR score <7, which occurred more frequently in the TOL group, was attributable to the lower Apgar scores in the TOL:vaginal group (Twin A aOR 6.4, 95%CI 1.5–27.4; Twin B aOR 6.1, 95%CI 1.7–21.4). The protective effect against TTN was only realized if TOL resulted in a vaginal delivery (Twin A aOR 0.4, 95%CI 0.2–0.6; Twin B aOR 0.3, 95%CI 0.2–0.6).

For twins admitted to the NICU, Twin A's NICU length of stay was 4.0 days (IQR 2.0–8.0 days) for the TOL:vaginal group, 4.0 days (IQR 3.0–8.5 days) for the TOL:caesarean group, and 4.3 days (IQR 3.0–8.5 days) for the PLC group, = 0.72. Twin B's NICU length of stay was 6.0 days (IQR 3.0–11.0 days) for the TOL:vaginal group, 4.0 days (IQR 3.0–8.0 days) for the TOL:caesarean group, and 5.0 days (IQR 2.0–8.0 days) for the PLC group, = 0.14.

For the combined vaginal-caesarean delivery group, the mean blood loss was 814 ml, range 150–2100 ml. Blood transfusions occurred in 7/57 (12.3%). The mean birthweight was 2710 g for Twin A and 2800 g for Twin B. Twin A neonatal outcomes included one birth injury, seven NICU admissions (12.3%), and an NICU mean length of stay of 3.6 days. Twin B neonatal outcomes included one birth injury, 11 NICU admissions (19.3%), one 5-minute APGAR <7, two cases of RDS, and NICU mean length of stay of 4.6 days. There was no asphyxia or neonatal deaths in this group.

Discussion

Main findings

In this cohort study of twins at term and their delivery mode, the majority of women attempting TOL delivered vaginally. The morbidities associated with TOL included postpartum haemorrhage, blood transfusions, birth trauma, and risk of 5-minute Apgar score <7. These outcomes were countered by less TTN in TOL. Overall, the burden of morbidity was greater for the maternal outcomes over the neonatal outcomes. Although neonatal findings were statistically significant, the overall occurrences of trauma and 5-minute Apgar score <7 were low.

Strengths and limitations

The strengths of this investigation include the large sample size, the scope of this multicenter study, and the number of maternal and neonatal outcomes available for analysis. To our knowledge, this is the largest study to evaluate the planned mode of delivery in term twins, rather than only the actual delivery mode. The large sample size also limits the interpretations when specific aspects of delivery events are not detailed. The decision for TOL vs PLC was at the discretion of the provider and likely varied among different sites; however, it is known that all the sites did allow twin trials of labour. As the TOL and PLC groups had dissimilar maternal characteristics, the effects of selection bias were mediated by controlling for confounders, such as prior caesarean. Finally, many of the outcomes occurred infrequently and a much larger cohort would be required to have the power to detect a significant difference if one truly exists.

Misclassification bias is also a potential source of bias relating to the creation of the TOL group. The definition of TOL used for this study was a vaginal delivery or having at least two cervical examinations documented in the database for that admission. It is possible that women who were to undergo a planned PLC, but who presented for a labour evaluation and therefore had two cervical exams, may be incorrectly classified to the TOL group. If these were women who experienced postpartum haemorrhage or blood transfusions, which were significant findings in the TOL:caesarean group compared with the PLC group, then the results would have been biased towards a significant finding that may not actually be significant, or a type 1 error. Alternatively, the magnitude of the misclassification could have been so great that the results may be significant in the opposite direction, with greater risk for haemorrhage and blood transfusions associated with PLC than with TOL. The inclusion criteria of the ‘term' twin likely lessened the number of labour evaluations for those with planned PLC at admission and thereby supports the correct classification of women. In addition, the percentage of twins undergoing a trial of labour was similar compared with other studies, as was the rate of vaginal deliveries.[17, 18, 21]

Interpretation

Two retrospective cohort studies evaluated outcomes based on planned mode of delivery in twin gestations, both of which employed protocols for active second-stage management. Schmitz et al.[18] evaluated 758 consecutive sets of twins >35 weeks with a cephalic presenting first twin, and controlled for pregnancy complications such as hypertension, pre-eclampsia, diabetes, twin-to-twin transfusion syndrome, and placenta praevia. Similar to our study, they also found increased postpartum haemorrhage in the planned vaginal delivery group, but no difference in transfusions between the groups.[18] They found no difference in neonatal composite morbidity for either twin when comparing outcomes in planned vaginal delivery to planned caesarean delivery. Fox et al. reported only neonatal outcomes in 287 twin gestations and found no difference in Apgar scores or umbilical artery pH<7.2 in the planned vaginal versus planned caesarean group.[17]

Both of the aforementioned study findings are consistent with the only randomized controlled trial of the management of a nonvertex second twin which randomized 60 twin gestations to vaginal delivery or caesarean delivery and found no significant differences in Apgar scores or other neonatal morbidity in the second twins.[12] In addition, a meta-analysis of four studies also noted no evidence to support planned caesarean delivery of twins, unless Twin A was breech.[11] However, many of the outcomes still occurred infrequently, making it difficult for the studies to have enough statistical power to detect a difference.

Finally, Peaceman et al. reported neonatal morbidities and death in the first year of life in 450 504 twin gestations from a Matched Multiple Birth dataset that compared twins delivering at ≥30 weeks by actual delivery mode for vertex/vertex and vertex/nonvertex pairs.[22] A sub-analysis of neonatal outcomes for infants born ≥34 weeks' gestation found lower Apgar scores and increased birth injuries as well as an increased overall composite neonatal morbidity score in the vaginal compared with the caesarean group, similar to the findings for the entire cohort ≥30 weeks' gestation. However, they concluded that these statistically significant differences were not clinically important and that neonatal morbidity and mortality were similar for vaginal and caesarean deliveries of twin gestations ≥30 weeks' gestation.

Combined vaginal-caesarean delivery occurred in 5.3% of our TOL cohort. This is similar to the 6–16% rate published in other reports, although the denominator in these publications may vary between a TOL group or a vaginal delivery of the first twin group.[23-25] This percentage may increase to 22–23% with vertex/nonvertex pairings after the first twin delivers vaginally.[22, 23] Our database was not able to delineate between our vertex/vertex and vertex/non-vertex pairings.

Conclusion

This study reflects on contemporary practices of twin delivery modes in the USA. In the entire cohort, 68.1% of all twins at term were delivered via caesarean. For those women who had a TOL, 65.9% had a vaginal delivery, a relatively high degree of success. In concordance with our original hypothesis we have demonstrated increased maternal and neonatal morbidity in the TOL group. Specifically, the maternal morbidity in the form of increased transfusions was related to caesarean delivery after a trial of labour. The neonatal morbidity, in the form of birth trauma and 5-minute Apgar score <7, was rare overall, and related to vaginal delivery rather than caesarean delivery after TOL, in contrast to our secondary hypothesis. It is important to note the absolute increase in birth asphyxia in the TOL group, although this finding was not statistically significant. These rare neonatal outcomes must be balanced against the lack of difference in NICU admissions and decreased TTN in the TOL group. Increased birth trauma suggests that vaginal delivery of twins may be a technically difficult procedure that requires a skilled provider. This study supports that a TOL is a reasonable and safe option that should be strongly considered after counselling women regarding the maternal and neonatal risks and benefits for twin gestations at term.

Disclosure of interests

The authors report no conflicts of interest.

Contribution to authorship

D.J.W. was the primary author of this paper and was responsible for design, acquisition of data, analysis and interpretation, writing and final approval. W.G. was responsible for aid in the design and analysis of data, revision of the methods section of the paper and final approval. M.A.K. contributed to the conception and design, acquisition of data, analysis and interpretation, critical revisions of the paper and final approval. I.W. contributed to the conception and design, critical revisions of the paper, and final approval.

Details of ethics approval

IRB #2007-0656, date of approval 9/22/2008.

Funding

This study was supported by the Intramural Research Program of the Eunice Kennedy Shriver National Institute of Child Health and Human Development, National Institutes of Health, through a contract (Contract No. HHSN267200603425C), by Grant Number K12HD055892 from the NICHD and NIH office of Research on Women's Health (ORWH), and the University of Illinois at Chicago (UIC) Center for Clinical and Translational Science (CCTS), Award Number UL1RR029879 from the National Center for Research Resources.

Acknowledgements

The data included in this paper were obtained from the Consortium on Safe Labor, which was supported by the Intramural Research Program of the Eunice Kennedy Shriver National Institute of Child Health and Human Development, National Institutes of Health, through Contract No. HHSN267200603425C. Institutions involved in the Consortium include, in alphabetical order: Baystate Medical Center, Springfield, MA; Cedars-Sinai Medical Center Burnes Allen Research Center, Los Angeles, CA; Christiana Care Health System, Newark, DE; Georgetown University Hospital, MedStar Health, Washington, DC; Indiana University Clarian Health, Indianapolis, IN; Intermountain Healthcare and the University of Utah, Salt Lake City, UT; Maimonides Medical Center, Brooklyn, NY; MetroHealth Medical Center, Cleveland, OH.; Summa Health System, Akron City Hospital, Akron, OH; The EMMES Corporation, Rockville MD (Data Coordinating Center); University of Illinois at Chicago, Chicago, IL; University of Miami, Miami, FL; and University of Texas Health Science Center at Houston, Houston, Texas. The named authors alone are responsible for the views expressed in this manuscript, which do not necessarily represent the decisions or the stated policy of the NICHD.

Ancillary