Methods	RCT 2-arm parallel group design.
Participants	Setting: (not clear) hospital in Oklahoma, USA. 100 women in early active labour (with regular contractions and cervical dilatation 3-4 cm); at term (at or > 37 weeks’ gestation); no medical or obstetric complications or evidence of fetal distress; requesting a “pain shot” rather than an epidural (all women were offered epidural)
Interventions	Both groups had continuous electronic fetal monitoring and intrauterine pressure catheters Experimental: IV fentanyl 50-100 mcg every 1-2 hrs to a max of 5 doses Control: IV butorphanol 1-2 mg every 1-2 hrs to a max 5 doses (Doses of drugs were approximately equivalent in both arms of the trial.)
Outcomes	Maternal uterine activity; adverse effects and side effects (including vomiting and sedation); pain scored using 10-point VAS (0 = no pain, 10 = excruciating pain) scores were recorded by nurses; Apgar scores at 1 and 5 min; infant neurological exam 2-4 and 24-36 hrs after birth
Notes
Risk of bias
Bias	Authors’ judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Computer-generated randomisation schedule.
Allocation concealment (selection bias)	Low risk	Pharmacy prepared identical unlabelled, coded syringes.
Blinding (performance bias and detection bias) Women	Low risk	Identical syringes.
Blinding (performance bias and detection bias) Clinical staff	Low risk	Described as double blind.
Blinding (performance bias and detection bias) Outcome assessor	Low risk	Outcome assessors reported as blinded.
Incomplete outcome data (attrition bias) All outcomes	Unclear risk	It was not clear at what point women were randomised. 155 women enrolled; 24 decided to have an epidural and were excluded (it was not clear whether or not this was after randomisation); 19 women delivered within one hour of first dose and 12 did not request analgesia and were not included in the analysis. Data available for 100 women; if loss occurred after randomisation this represents a very high level of attrition
Selective reporting (reporting bias)	Unclear risk	Unclear.
Other bias	Low risk	None apparent.