Methods	RCT with randomisation of individual women.
Participants	Irish hospital setting. High-income country. Inclusion criteria: all women at low risk of haemorrhage (< 35 years; parity < 5; 1st stage of labour <15 hours; no previous history of PPH; Hb >11 g/dL (or 10.6 g/dL for capillary sample)) with singleton, cephalic presentation, 35 to 36 weeks at recruitment; no medical complications which would contraindicate ergometrine or would increase the risk of bleeding (cardiac disease, use of heparin, hypertension), and expected to give birth vaginally 1429 women randomised out of 2901 eligible. Exclusion criteria: women with hypertension in pregnancy or 1st or 2nd stage; epidural anaesthesia (included in separate study); APH; 1st stage >15 hours; OVB; women attending private care Clinician responsible for third stage: midwives.
Interventions	Intervention: active management of third stage (N = 705): prophylactic ergometrine 0.5 mg IV immediately following birth (once 2nd twin excluded); try to clamp cord within 30 sec; delivery of the placenta by controlled cord traction (CCT) when uterus contracted; try not to give any special instructions re posture. For retained placenta: 1 hour after birth: ensure empty bladder; attempt delivery again using CCT; manual removal of placenta under GA. Comparison: expectant management of third stage (N = 724): no oxytocic drug to be given routinely; try to leave cord attached to baby until pulsation has ceased. When cord is cut milk any placental blood into bowl and discard; encourage mother to breastfeed; watch for signs of placental separation and ask women to tell you when she feels a contraction or urge to push. DO NOT touch the abdomen or manipulate the uterus at this stage; placenta may be delivered by maternal effort or gentle controlled cord traction; if mother does not experience a contraction within 8 minutes, place hand gently on fundus to determine intrauterine bleeding and await separation. Special circumstances: if baby’s cord is clamped and cut before pulsation ceases (due to cord round neck, asphyxia etc) do not give ergometrine. Milk any placental blood into bowl and discard it. Watch for signs of placental separation and deliver placenta by controlled cord traction Retained placenta > 1 hour after birth: ensure empty bladder; attempt delivery again using controlled cord traction; give ergometrine 0.5 mg IV and re-attempt delivery; manual removal under GA. Data entered into Comparisons 1 and 2
Outcomes	Pre-specified outcomes: manual removal of placenta; PPH (> 500 mL); mean blood loss; length 3rd stage; Hb <10 g/dL at 48-72 hours; and difference between 32 weeks and 4872 hours PP: PP blood transfusions; side effects 1-2 hours post birth; PP complications; breastfeeding; serum prolactin; women’s views. No neonatal outcomes (Information from Oxford Database of Perinatal Trials registration sheet) and from [Begley 1990]): morbidity; blood loss during 3rd stage; method of placental delivery; complications occurring in first 1-2 hours post birth (haemorrhage, nausea, vomiting, raised BP, pain); Hb on 3rd postnatal day; prolactin levels on 3rd postnatal day, duration of breastfeeding.
Notes	Between 1 Oct ‘87 and 31 Oct ‘88, 2901 women were deemed eligible for initial inclusion, 2650 agreed to take part. 1221 of these were excluded prior to randomisation because of epidural (399); OVB (354); CS (132); rapid birth (95); hypertension (77); missed (53); low Hb (40); woman’s request (28); miscellaneous (23); breech (20) Actual management used in the active arm: all given IV ergometrine 0.5 mg before delivery of placenta; 89% cord clamped and cut; 93% CCT and 5% maternal effort; 7% upright and 93% recumbent Actual management used in the expectant arm: 14% got ergometrine for treatment, not prophylactically, 6 (0.83%) before placenta delivered; cord left unclamped till pulsation ceased 42%; placenta delivered by maternal effort 32% and CCT 66%; 11% upright
Risk of bias
Bias	Authors’ judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	Unpublished information from author: random number tables were used (Fleiss 1981). The first number was selected from the table by a disinterested observer and the numbers were allocated in blocks of 100 following in sequence
Allocation concealment (selection bias)	Low risk	“…a numbered, sealed envelope containing the randomly allocated group was stapled to the woman’s chart in readiness for admis-sion…The envelope remained sealed until the women was in second stage of labour and the midwife was certain a normal delivery would ensue. The envelope was then opened…”
		“When a woman was excluded from the study, her envelope was returned, unopened, to the researcher. All returned envelopes were re-allocated in numerical order prior to starting the next batch of 100 envelopes.”
Incomplete outcome data (attrition bias) All outcomes	Low risk	Some missing data for some postnatal haemoglobin measurements (618 out of 705 in the active group (12% attrition) and 645 out of 724 in the expectant (11% attrition) ITT not mentioned but no loss to follow up for outcomes measured during labour. Some outcome data are taken from the unpublished thesis
Selective reporting (reporting bias)	Low risk	None apparent, outcomes on trial registration sheet all reported
Other bias	Low risk	No significant difference in baseline characteristics, but more women in the physiological arm had pethidine in labour (46% cf 52%, P = 0.05). This may impact on outcomes in the physiological arm where the sight and sound of the baby may be the stimulus for the hormonal release needed for natural 3rd stage and pethidine may impact here. 1st and 2nd stage management similar and no obvious differences overall