Gulmezoglu 2001.

Methods	2‐arm active‐controlled double‐blinded randomised trial
Participants	18,530 women were randomised in a hospital setting in Argentina, China, Egypt, Ireland, Nigeria, South Africa, Switzerland, Thailand and Vietnam Nigeria, South Africa, Switzerland, Thailand, and Vietnam. The population comprised women of both nulliparous and multiparous, unspecified whether singleton or multiple pregnancy, at both high and low risk for PPH, who delivered by vaginal delivery. Exclusion criteria comprised women undergoing elective or emergency caesarean section after randomisation, or those with asthma, severe chronic allergic conditions, abortion, pyrexia (more than 38°C) or inability to give consent.
Interventions	600 mcg of misoprostol administered orally versus 10 IU of oxytocin administered IM or by an IV bolus
Outcomes	The study recorded the following outcomes: PPH at 500; PPH at 1000. Severe maternal morbidity; intensive care admissions; additional uterotonics; transfusion; manual removal of placenta; death; blood loss (mL); third stage duration (minutes); diarrhoea; nausea; vomiting; fever; shivering.
Notes	Contact with study authors for additional information: yes. Additional data from authors: yes
Risk of bias
Bias	Authors' judgement	Support for judgement
Random sequence generation (selection bias)	Low risk	The random allocation schedule was generated centrally at WHO, Geneva, Switzerland, by computer‐generated random numbers and was stratified by country. Within the strata, women were individually randomised into 1 of 2 intervention groups with randomly varying block sizes of 4–6 women.
Allocation concealment (selection bias)	Low risk	The treatment packs were sealed, numbered sequentially, and could only be taken from the dispenser consecutively.
Blinding of participants and personnel (performance bias) All outcomes	Low risk	Quote: "The treatment packs and their contents were identical in shape, colour, weight, and feel."
Blinding of outcome assessment (detection bias) All outcomes	Low risk	Assessors were blinded to treatment allocations.
Objective assessment of blood loss	Low risk	Investigators appraised blood loss from the time of delivery of the baby until the third stage of the labour was completed, when the mother was transferred to postnatal care (usually up to 1 hour postpartum). Immediately after the cord was clamped and cut, they passed a flat bedpan or an unsoiled receiver under the mother. The collected blood was poured into a standard measuring jar provided by WHO for volumetric measurement. Quote: "To simplify the procedure... small gauze swabs soaked with blood were put into the measuring jar and included in the measurement together with the blood and clots".
Incomplete outcome data (attrition bias) All outcomes	Low risk	Investigators excluded quote: "37 and 34 women with emergency caesarean section, and 13 and 4 women lost to follow‐up in misoprostol and oxytocin groups, respectively, for blood loss ≥ 1000 mL, and 2 and 4 women without information on the need for additional uterotonics".
Selective reporting (reporting bias)	Low risk	The study report matches the study protocol that was published in advance.
Intention to treat analysis	High risk	Not all study participants were included in the analysis.
Funding source	Low risk	The study was supported by funding from the UNDP/UNFPA/WHO/World Bank (public funding). Special Programme of Research, Development and Research Training Human Reproduction of WHO. Searle (Skokie, IL, USA) and Novartis (Basel, Switzerland) donated the active and placebo medications used in the trial.