Bollapragada 2009.

Methods	Double‐blind randomised controlled trial in an outpatient setting. Recruitment between March 2005 and December 2006 at Princess Royal Maternity Hospital, Glasgow.
Participants	Women scheduled for admission for cervical ripening and labour induction. Inclusion criteria: included all of the following: nulliparity, singleton fetus, cephalic presentation, ≥ 37 completed weeks' gestation, modified Bishop score < 7, and willingness to self‐administer vaginal tablets. Exclusion criteria: included women < 16 years of age, those who needed delivery within the next 48 hours in the fetal or maternal interest or who had ruptured membranes.
Interventions	350 were randomised. 177 were prescribed 40 mg ISMN tablets and 173 received placebo, with instructions to self‐administer the tablets vaginally at home at 48, 32 and 16 hours prior to scheduled time of admission. After admission to hospital, induction of labour was with vaginal prostaglandins until cervical ripening (described as Bishop score > 6) was achieved or 3 doses of prostaglandin tablets (3 mg each) were administered. Once the cervix was ripe fetal membranes were ruptured and oxytocin administered if required.
Outcomes	Maternal: elapsed time from admission to delivery, operative delivery rates (caesarean section and instrumental vaginal delivery), vaginal delivery not achieved in 24 hours, cervix unfavourable/unchanged at 12 to 24 hours, oxytocin augmentation, epidural analgesia, maternal side effects, postpartum haemorrhage, requirement for additional inpatient cervical ripening agents. various outcomes relating to maternal satisfaction. Neonatal: serious neonatal morbidity/perinatal death, meconium‐stained liquor, admission and duration of NICU admission, 5‐minute Apgar score of less than 7.
Notes	Detailed economic data also included. Protocol published previously.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Centrally‐generated randomisation schedule in permuted blocks of 4.
Allocation concealment (selection bias)	Low risk	Pharmacy at Western Infirmary in Glasgow prepared identical treatment packs, labelled with relevant unique study number. Allocation via automated interactive telephone response service.
Incomplete outcome data (attrition bias) All outcomes	Low risk	47 in ISMN arm and 46 in placebo arm withdrawn after randomisation. Majority went into spontaneous labour. 11 withdrawals in total, 2 diagnosed with breech presentations and hence excluded.
Selective reporting (reporting bias)	Low risk	No evidence to the contrary.
Other bias	Low risk	No evidence to the contrary.
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Patients, therapists, outcome assessors and analysts blinded to allocation.
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Patients, therapists, outcome assessors and analysts blinded to allocation.