Gestational age profile for spontaneous birth in twin and triplet pregnancies

One study¹⁶³ was identified for inclusion in relation to incidence of spontaneous birth in twin and triplet pregnancies by gestational age. This study reported data from all twin births in New South Wales, Australia, for a period of 10 years (1990–1999). No study reporting similar data for spontaneous birth in triplet pregnancies was identified.

Perinatal mortality and morbidity in spontaneous or uncomplicated birth at different gestational ages

No studies were identified for inclusion in relation to perinatal outcomes of spontaneous birth in uncomplicated twin and triplet pregnancies according to gestational age at birth. Two large, population-based studies from Japan and the UK, ^164;165 reported data on fetal death rates according to gestational age in multiple (predominantly twin) pregnancies. These studies did not make any distinction between monochorionic and dichorionic twin pregnancies.

To explore the effect of chorionicity on fetal death rates at different gestational ages, six smaller studies reporting data for monochorionic twin pregnancies were identified.^166–171 Three of the studies reported data for dichorionic twin pregnancies in the same population.^166–168

Neonatal mortality among twins born at different gestational ages was reported in three studies.^164;166;170 A Japanese study reported neonatal morbidity according to gestational age at birth in dichorionic twins.¹⁷²

Two small studies reported fetal death rates at different gestational ages in triplet pregnancies.^173;174 In addition, one large study examined US data over 4 years and compared twin and triplet pregnancies to singleton pregnancies for stillbirth and neonatal mortality rates specific to gestational age.¹⁵⁹

Effectiveness of elective delivery in twin and triplet pregnancies

Three studies, including one RCT,¹⁷⁵ one quasi-randomised trial¹⁷⁶ and one retrospective observational study¹⁷⁷ were identified that compared elective delivery with expectant management in twin pregnancies. No studies were identified that compared elective delivery with expectant management in triplet pregnancies.

Gestational age profile for spontaneous birth in twin and triplet pregnancies

One cross-sectional study (low quality evidence) suggested that the majority (58%) of women with uncomplicated twin pregnancies give birth spontaneously before 37 weeks 0 days. No robust data were identified for the gestational age profile in spontaneous triplet births.

Perinatal mortality and morbidity in spontaneous or uncomplicated birth at different gestational ages

No evidence was identified for perinatal outcomes at different gestational ages in uncomplicated twin or triplet pregnancies with spontaneous onset of labour.

Indirect evidence from studies reporting all multiple pregnancies together (uncomplicated, spontaneous onset of labour or otherwise) demonstrated an increase in the risk of fetal death per week towards the end of pregnancy. In the largest study of multiple pregnancies (predominantly dichorionic twin pregnancies), the relative risk of fetal death per week of gestation compared to the risk in singleton pregnancies at 42 or more weeks of gestation rose significantly from 37 weeks (low quality evidence). In the same study, early neonatal mortality (death within 7 days of live birth) showed similar trends, with the lowest death rate being reported in twins born at 37 weeks of gestation (low quality evidence).

Most studies that reported fetal or neonatal death rates separately for different types of multiple pregnancy (monochorionic twin, dichorionic twin or triplet pregnancies) were underpowered to detect differences in death rates between clinically important gestational ages (for example between fetal death rates at a given gestational age and those at 37 weeks 0 days to 37 weeks 6 days). The fetal death rate in monochorionic twin pregnancies was significantly higher than that in dichorionic twin pregnancies at 36 weeks or more, and point estimates of relative risk of fetal death were greater than one at all gestational ages (very low quality evidence). In monochorionic twin pregnancies and dichorionic twin pregnancies, fetal death rates were consistently lower at gestational ages between 26 weeks and 35 weeks compared to 36 weeks or more, although not significantly lower (very low quality evidence). In triplet pregnancies, fetal death rates were consistently lower at 33 weeks to 36 weeks compared with 37 weeks or more, and significantly lower at 34 weeks (very low quality evidence). In monochorionic twin pregnancies, neonatal death rates were significantly higher at gestational ages up to 29 weeks compared with 37 weeks or more, and the rates declined further from 30 to 35 weeks (very low quality evidence). In one study involving triplet pregnancies, no serious neonatal morbidity (respiratory distress syndrome, chronic lung disease, intraventricular haemorrhage grades 3 or 4, necrotising enterocolitis or proliferative retinopathy of prematurity) was reported after 34 weeks (very low quality evidence).

Effectiveness of elective delivery in twin and triplet pregnancies

Three studies showed no clinically significant difference in neonatal or maternal outcomes between women with twin pregnancies who underwent elective delivery and those who underwent expectant management (low to high quality evidence).

No studies were identified that examined effectiveness of elective delivery in women with triplet pregnancies.

Relative value placed on the outcomes considered

The following were considered to be critical outcomes for this review question:

• perinatal mortality, neonatal mortality or stillbirth
• neonatal respiratory problems
• admission to a neonatal unit
• neonatal encephalopathy
• maternal morbidity (such as postpartum haemorrhage requiring blood transfusion, hypertension)
• operative delivery (instrumental delivery or caesarean section)
• Apgar score
• birthweight.

Trade-off between clinical benefits and harms

The evidence reviewed by the GDG indicated that 58% of women with twin pregnancies give birth spontaneously before 37 weeks 0 days. No comparable evidence was identified for triplet pregnancies; however, the GDG is aware of literature suggesting that about 75% of women with triplet pregnancies give birth spontaneously before 35 weeks 0 days.¹⁷⁹ The baseline risks of spontaneous preterm birth and its consequences, especially for babies, and the comparative risks of fetal death at increasing gestational ages are the main focus of attention in this review question, which seeks to identify the optimal timing of birth for women with twin and triplet pregnancies. For twin pregnancies the main clinical harm is the increasing risk of fetal death towards the end of pregnancy; this appears to be disproportionately greater in monochorionic twin pregnancies. Hence the GDG’s view was that women with dichorionic twin pregnancies should be offered elective birth at 37 weeks 0 days, whereas those with monochorionic twins should be offered elective birth at 36 weeks 0 days.

For triplets there are two clinical risks with continuing pregnancy towards the end of the third trimester. One is the risk of spontaneous preterm labour and delivery occurring in an adverse setting, the other is a significantly higher risk of fetal death after 34 weeks 6 days. Thus, the GDG’s view was that women with triplet pregnancies should be offered birth at 35 weeks 0 days.

The main trade-offs between clinical benefits and harms for women with twin and triplet pregnancies who have not given birth spontaneously at a given gestational age are the risks of neonatal mortality and morbidity or maternal operative delivery associated with elective delivery versus the risks of fetal death (stillbirth) from continued pregnancy. The GDG acknowledged that the evidence regarding neonatal morbidity associated with elective birth in twin and triplet pregnancies was limited and further research is needed.

It would be helpful in clinical practice to inform women of the absolute risks of fetal death in twin and triplet pregnancies. While this is possible for twin pregnancies (using the fetal death rate per 1000 fetuses for a given gestational period), it is not possible for triplet pregnancies (because the rates are available only as fetal deaths per 1000 live births). The GDG’s view is that it would be confusing to quote absolute fetal death rates for twin and triplet pregnancies in different units. This is why the GDG’s recommendations do not include absolute fetal death rates.

Trade-off between net health benefits and resource use

The health economic analysis conducted for this review question showed that prolonging twin pregnancies beyond 37 weeks 0 days and triplet pregnancies beyond 35 weeks 0 days would incur the loss of health benefits (QALYs), albeit at an increasing cost, and this would not represent value for money. To maximise health benefits in uncomplicated twin and triplet pregnancies, birth should be at 36 weeks 0 days in dichorionic pregnancies, 37 weeks 0 days in monochorionic twin pregnancies and 35 weeks 0 days in triplet pregnancies. This is expected to result in cost savings to the NHS. The GDG recognised that it may be appropriate to offer birth even earlier than 37 weeks 0 days, 36 weeks 0 days, or 35 weeks 0 days if clinically indicated.

Quality of evidence

The evidence ranged in quality from very low to moderate. The best available evidence was sufficient to demonstrate that elective birth by 37 weeks 0 days would be cost effective for all types of multiple pregnancy. Observational studies that reported fetal or neonatal death rates separately for different types of multiple pregnancy were underpowered to detect differences in death rates between clinically important gestational ages (for example, between fetal death rates at a given gestational age and those at 37 weeks 0 days to 37 weeks 6 days), and so recommendations for elective birth at 36 weeks 0 days in monochorionic twin pregnancies and 35 weeks 0 days in triplet pregnancies incorporated consideration of current practice in addition to the available evidence. Further research is needed to determine precisely the optimal timing of birth according to chorionicity and multiplicity of the pregnancy.

Other considerations

The GDG recognised the importance of offering antenatal administration of corticosteroids for elective preterm birth in monochorionic twin pregnancies and triplet pregnancies. The specialist team should discuss with all women with twin and triplet pregnancies the possibility of their babies being admitted to a special care unit if they have a spontaneous preterm birth or if the offer of elective preterm birth is accepted. The GDG also recognised the importance of ensuring that ongoing care is provided for women with twin and triplet pregnancies who decline the offer of elective early birth. No evidence was identified in relation to the optimal surveillance strategy for pregnancies that continue beyond 37 weeks 0 days, 36 weeks 0 days or 35 weeks 0 days in dichorionic twins, monochorionic twins and triplets, respectively. The GDG’s recommendation for weekly appointments with the specialist obstetrician, with weekly biophysical profile testing of all fetuses and fortnightly growth scans, was based on the GDG members’ collective experience.

The possibilities for elective birth are induction of labour or caesarean section. Consideration of mode of delivery is outside the scope of this guideline (because it relates to intrapartum care, not antenatal care), although the GDG was aware that in triplet pregnancies, for example, caesarean section is currently used more frequently than induction of labour. The footnotes to the recommendations relating to timing of birth emphasise that mode of delivery is outside the scope of the guideline.

The GDG highlighted the importance of a member of the core team starting discussions and planning regarding timing of birth and mode of delivery before the time at which elective birth would occur if the offer was accepted.

Recommendations