Impey 2005.

Methods	RCT.
Participants	Inclusion criteria: nulliparous women, singleton breech presentation at 36 weeks, or multiparous at 37 or more weeks. Eligible for inclusion if an unsuccessful attempt at ECV (without tocolysis) was reported, with normal CTG. N = 144. Exclusion criteria: preexisting indication for CS, unstable lie, fetal compromise (abdominal circumference below 3rd centile, either umbilical artery resistance index above 97th centile or deepest amniotic fluid pocket 2 cm, rhesus isoimmunisation.
Interventions	Intervention:tocolytic: ritodrine ‐ beta stimulant (A1) ‐ parenteral. Tocolysis administered as ritodrine hydrochloride (Yutopar infusion of 50 mg (10 mg/mL)) added to 12 mL dextrose saline (total 17 mL of ritodrine 3 mg/mL). N = 62. Comparison:placebo: 17 mL dextrose saline infusion by the same route at the same rate. N = 62.
Outcomes	Primary outcome cephalic presentation at birth. Secondary outcomes: incidence of successful ECV, CS, length of hospital inpatient stay, incidence of neonatal Apgar scores < 7 at 5 minutes, neonatal admission, rare neonatal outcomes and mean cord arterial pH. McGill pain scale was used to measure intensity of pain.
Notes	Setting: Breech Clinic, John Radcliffe Hospital, Oxford: women from community clinics and other local hospitals were referred in at 36 or 37 weeks' gestation.
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Randomly allocated in a ratio of 1:1 using random block sizes up to 20.
Allocation concealment (selection bias)	Low risk	Sealed opaque envelopes opened in sequential order.
Blinding of participants and personnel (performance bias) All outcomes	Unclear risk	Specific detail missing, but states same infusion, same timeline, same observation for both control and intervention. In discussion, study authors identified problems with blinding of researcher and medical practitioner as potential threats.
Blinding of outcome assessment (detection bias) All outcomes	Unclear risk	Specific detail missing, but states same infusion, same timeline, same observation for both control and intervention. In discussion, study authors identified problems with blinding of researcher and medical practitioner as potential threats.
Incomplete outcome data (attrition bias) All outcomes	Low risk	Potential sample 505, of whom 284 were deemed eligible for inclusion in the  trial. Of these, 13 refused and 47 were not offered. All 124 participants (62 in each arm) completed the trial. Intention‐to‐treat analysis.
Selective reporting (reporting bias)	Unclear risk	We did not assess the trial protocol.
Other bias	Low risk	No differences in age or gestation seen in baseline data.