Twin pregnancy presents numerous challenges for the obstetrician from conception onwards, none more so than the timing and mode of delivery.

In the majority of cases, the timing of delivery is not subject to the considered appraisal of the obstetrician, since spontaneous preterm birth complicates between 30 per cent and 50 per cent of cases¹ and intrauterine growth restriction (IUGR) further complicates about one-third of twin pregnancies². In addition, major congenital anomalies may affect up to 4.9 per cent of twin pregnancies. To top things off, there are several fetal problems specific to twin pregnancy, including conjoined twins, twin-reversed arterial perfusion (TRAP) sequence, monoamnionic twinning and twin-to-twin transfusion syndrome (TTTS). All of these require further careful consideration with regard to planning of the timing and mode of delivery in discussions between patient and obstetrician. This article, however, will concentrate on the timing and mode of delivery of  ‘uncomplicated’ twins near term.

A retrospective study by Dodd and colleagues³ suggested that the lowest rate of perinatal mortality and morbidity in twin pregnancies occurs with delivery between 36 and 38 weeks gestation. The risk of adverse outcomes increases with advancing gestation, but a Cochrane Systematic Review comparing elective birth at 37 weeks gestation with continued expectant management failed to identify any statistically significant difference in the outcomes between the two approaches. A multicentre randomised controlled trial, coordinated by the University of Adelaide, is in progress and aims to assess the optimal timing of birth in women with a twin pregnancy at term.⁴ Until such information is available, there is a general consensus that the timing of delivery of uncomplicated twins should be between 37 and 38 weeks gestation.

The optimal mode of delivery of uncomplicated twins is controversial. There is non-randomised evidence suggesting that, for triplets and higher order multiple pregnancies, the risk of low Apgar scores and perinatal death is reduced with caesarean section.⁵ The evidence is less clear for twins, hence the obstetrician’s dilemma.

Smith and colleagues’ retrospective cohort study suggested that planned caesarean section may reduce the risk of perinatal death of twins (particularly the second twin) at term compared with attempting vaginal birth, but after correction for confounding factors, the conclusions were less robust.⁶

Chasen’s group reported that neonatal respiratory disease was more common in twins born by caesarean section at 36 to 38 weeks compared to those born vaginally at 38 to 40 weeks.⁷ Also, Wildschut and colleagues reported increased perinatal mortality rates associated with caesarean delivery primarily as a result of neonatal respiratory distress.⁸

A Cochrane review by Crowther⁹, reviewing the mode of birth of the second twin, identified a single randomised trial by Rabinovici and colleagues¹⁰ comparing planned vaginal birth with planned caesarean birth for the second non-vertex twin, with the unsatisfactory conclusion that further evidence was required from randomised trials to determine the optimal mode of delivery.

At the present time, a randomised controlled trial – the Twin Birth Study, coordinated by the University of Toronto – is underway. This study aims to provide more reliable information as to the optimal mode of delivery of uncomplicated twins.¹¹ While this study is in progress and information is obtained, it seems to be reasonable clinical practice to offer the patient with uncomplicated twins the choice of elective caesarean section or selected elective vaginal delivery at 37 to 38 weeks gestation.

With regard to the choice of elective caesarean delivery versus vaginal delivery, based on the literature from the 20th century, certain recommendations can be made to obstetricians practising in the 21st century. Firstly, the mode of delivery may be affected by the respective presentation of the twins. The most common presentation of twins is vertex-vertex. In most cases, obstetricians recommend vaginal birth¹² and the available literature supports this option.¹³

With the first twin in vertex presentation and the second twin non-vertex, opinion is divided as to the optimal mode of delivery. Some obstetricians recommend elective caesarean section and report reduced neonatal mortality and morbidity rates.^13,14 Others suggest there is no increase in neonatal risk, even with external cephalic version and/or internal podalic version of the second twin at the time of delivery¹⁵, placing an emphasis on the reduction of risks for the mother.¹⁶

When the first twin is non-vertex, caesarean section is usually advised¹⁷, but in the case of breech presentation of the first twin, unfortunately, the Term Breech Trial¹⁸, which reported data for singleton births, seems to influence the mode of delivery offered, with the avoidance of the rare occurrence of twin entrapment by interlocking chins or heads being an additional advantage cited for the option of caesarean section.¹⁹

During labour and delivery of twins, the outcome will be enhanced by delivery in hospital and, if possible, in a tertiary or level II facility. Sensible precautions include intravenous access, cross-matching of blood, continous electronic fetal monitoring of both twins, epidural anaesthesia, additional midwifery and paediatric assistance, oxytocin infusion for uterine inertia (particularly after delivery of the first twin), and prophylactic oxytocin infusion after delivery to reduce the risk of primary postpartum haemorrhage. I would emphasise, though, that the quality of evidence for these intrapartum strategies is generally level III and the strength of the recommendations is either B or GPP.

It is recognised that perinatal mortality in twin pregnancies is five to ten times higher²⁰ and the perinatal morbidity, particularly from cerebral palsy, is eight times higher²¹ in twin pregnancies compared to singleton pregnancies. This increased risk of perinatal mortality and morbidity mostly results from factors unrelated to the mode of delivery, such as antepartum stillbirth, the effects of prematurity, congenital anomalies and chorionicity-dependent complications.

The other important factor that must be considered when assessing perinatal outcomes in twin pregnancy is that women with a multiple pregnancy have a two-fold increase in the risk of death compared with women with a singleton pregnancy.²²

It is widely recognised that in contemporary practice, approximately 60 per cent of twins (and in some selected centres almost 100 per cent of twins) are delivered by caesarean section, and this practice seems to be increasing.^23,24 This high rate of caesarean birth may reflect some obstetricians’ anxiety about vaginal twin delivery, which in turn may be evidence of defensive practice.

It is worth noting that patients’ demand for caesarean delivery may follow their appraisal of the literature, as discovered on the internet. The effect of patient-initiated internet searching, a phenomenon that is common, is increasingly impacting on all aspects of medical care, but particularly in obstetrics.

The practical difficulty in attempting to obtain high-quality evidence on which to base clinical management is that, in obstetrics, serious adverse outcomes are relatively infrequent. For example, delivery-related perinatal death of the second twin at term affects approximately one in 287 twin births. As a consequence, to appropriately power a relevant randomised trial would require study of approximately 6500 twin pregnancies to determine whether caesarean section would reduce the risk of perinatal death, an ambition that will never be met.^13,14

The lack of data presents obstetricians with a dilemma in making clinical decisions that are based on robust evidence, particularly with regard to the mode of delivery of uncomplicated twins at term. For the moment, while new data are awaited, responsible obstetricians can legitimately present to their patients a choice between elective caesarean section and assisted vaginal birth at 37 to 38 weeks gestation.