Methods

Randomised controlled trial.

Participants

Setting: recruitment between Dec 2003‐April 2011, 2804 women randomised in 106 centres in 25 countries. Participating centres assessed fetal growth and well‐being with the use of ultrasonography at least every 4 weeks and with the use of non‐stress or biophysical profile tests twice weekly if needed; were prepared to perform a CS within 30 minutes if necessary; and had anaesthetic, obstetrical, and nursing staff available in the hospital at the time of planned vaginal delivery.

Inclusion criteria: women with twin pregnancy between 32 weeks’ and 38 weeks 6 days gestation. First twin cephalic, both twins alive and weight estimated between 1500 g and 4000 g confirmed by ultrasound.

Exclusion: monoamniotic twins, fetal reduction at 13 or more weeks, lethal anomaly, contraindication to VB or previous participation in the twin birth study.

Interventions

Experimental intervention: 1398 women (2795 fetuses) randomised to planned lower segment caesarean delivery. If the first twin delivered vaginally in this group, then CS was attempted with the second twin if this was feasible.

Control/comparison intervention: 1406 women (2812 fetuses) randomised to planned vaginal delivery with induction of labour between 37 weeks 5 days and 38 weeks 6 days.

(Use of continuous electronic monitoring, epidural and oxytocin at the discretion of the obstetrician. Women were attended by obstetricians experienced in managing vaginal delivery of twins.)

Outcomes

Primary outcome: fetal or neonatal mortality or serious neonatal morbidity. (Neonatal morbidity 0‐27 days) Serious morbidity defined.

Maternal death or serious maternal morbidity before 28 days postpartum (defined).

Secondary outcomes, infant or child death or poor neurodevelopmental outcome up to 2 years. Maternal satisfaction with mode of delivery, breastfeeding, quality of life, fatigue or depression.

Notes

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Low risk

Centrally‐controlled randomisation by computer with stratification for parity and gestational age using random block sizes.

Allocation concealment (selection bias)

Low risk

See above. Centralised randomisation.

Blinding of participants and personnel (performance bias)
All outcomes

High risk

Blinding of women and staff is not feasible for this type of intervention. Lack of blinding could possibly lead to changes in management that might affect outcomes.

Blinding of outcome assessment (detection bias)
All outcomes

Unclear risk

There was an attempt to blind assessors for the primary outcome. It was not clear if this was successful.

Incomplete outcome data (attrition bias)
All outcomes

Low risk

There was very limited loss to follow‐up for the primary outcomes (2783/ 2795 fetuses included in analysis of primary outcome in the intervention group and 2782/2812 in the control group). Loss balanced in intervention and control groups. There was some further loss to follow‐up for longer‐term maternal outcomes.

ITT analysis.

Selective reporting (reporting bias)

Low risk

Protocol available as supplementary information and all outcomes appeared to be reported.

Other bias

Low risk

There was no evidence of baseline imbalance.

The analysis took account of correlation between outcomes for twins.

Methods

'Randomized control trial.'

Participants

Inclusion criteria: twin pregnancy, induced or spontaneous labour, both alive. 1st vertex, 2nd breech or transverse, 35‐42 weeks estimated gestational age, no known fetal anomaly, no signs of acute placental insufficiency or abruption, normal amniotic fluid volume, normal FHR testing, no maternal or obstetric indications for a specific route of delivery, cervix < 7 cm dilated. 60 women randomised.

Interventions

Experimental intervention: planned lower segment CS, preferably with epidural analgesia, but dependent on preference of anaesthetist.

Control/comparison intervention: planned VB following evaluation of labour progress using ‘Friedman curve’; continuous electronic fetal monitoring of both babies; lie of 2nd twin confirmed clinically or with ultrasound; if breech, assisted breech delivery planned; of fetal distress or poor progress despite oxytocin, total breech extraction done; artificial rupture of second sac as late as possible; if second twin in oblique or transverse lie, internal version and complete breech extraction under general analgesia or epidural analgesia if in place; routine episiotomy.

Outcomes

Method of delivery; birthweight, Apgar scores, neonatal and maternal complications.

Notes

Setting: Chaim Sheba Medical Centre labour ward.

Risk of bias

Bias

Authors' judgement

Support for judgement

Random sequence generation (selection bias)

Unclear risk

Allocation “changed randomly by a non‐involved person without prior notice on a time basis”.  20% difference in group sizes not accounted for (27 vs 33).

Allocation concealment (selection bias)

High risk

See above.

Blinding of participants and personnel (performance bias)
All outcomes

High risk

Blinding not feasible.

Blinding of outcome assessment (detection bias)
All outcomes

Unclear risk

Not mentioned whether neonatal assessments blinded.

Incomplete outcome data (attrition bias)
All outcomes

High risk

6 women allocated to planned VB excluded from primary analysis for delivery not according to protocol (2 CS and 4 vertex VBs). Analysis was not conducted on an ITT basis. Only categorical neonatal data have been included in the review as data given separately for the 6 excluded women and could be added to the primary data, thus low risk of bias for these outcomes.

Selective reporting (reporting bias)

Unclear risk

No pre‐published protocol available to check predefined outcome reporting.

Other bias

High risk

Baseline imbalance:  CS n = 27 vs vaginal n = 33.