Maternal mortality: one death every 7 min.

99% of all deaths in childbirth are in the least developed countries. Annually, 45 million women deliver without a skilled birth attendant, a situation in which the greatest number of maternal deaths occur. In many low-resource settings, having enough skilled birth attendants remains a distant goal. The commonest single cause of maternal death is from post-partum haemorrhage, from which one women dies every 7 min. The report by Mariana Widmer and colleagues in The Lancet today is a good example of the type of highquality clinical research being done in well-resourced hospitals. In this large trial (in more than 1400 women), the investigators used 600 μg sublingual misoprostol to treat clinically diagnosed (blood loss ≥500 mL) postpartum haemorrhage, after routine use of oxytocin in the third stage of labour. The study did not detect a diff erence between misoprostol and placebo. Today’s study complements two trials published earlier this year by the same group. One study found that 800 μg sublingual misoprostol and intravenous oxytocin were equally eff ective in controlling postpartum haemorrhage in women who had not received oxytocin in the third stage of labour, although the total amount of blood loss was less in the oxytocin group. The second found that 800 μg sublingual misoprostol was clinically equivalent to oxytocin in stopping postpartum haemorrhage in women who had already received prophylactic oxytocin and who were suspected of having uterine atony. These large, well-conducted randomised trials are part of an ongoing eff ort to build a solid and unambiguous evidence base to further improve obstetric care in well-resourced settings. However, as today’s paper concludes, “Any further research on misoprostol should focus on the possible eff ectiveness of misoprostol in settings where standard uterotonics are not available.” Here, an increasingly polarised debate is taking place about the nature of scientifi c evidence and clinical guidelines in the treatment of post-partum haemorrhage. For prevention of post-partum haemorrhage, pregnant women delivering at home without a skilled birth attendant can self-administer 600 μg misoprostol orally as soon as possible after their baby is delivered. Studies in Afghanistan, Nepal, and Bangladesh show that women use misoprostol consistently and safely (even for twin deliveries) when the drug is distributed at the community level. Misoprostol 1000 μg rectally has also been used successfully by traditional birth attendants (village midwives) to treat post-partum haemorrhage. In 2006, the International Federation of Gynecology and Obstetrics (FIGO) and the International Confederation of Midwives agreed that “in home births without a skilled attendant, misoprostol may be the only technology available to control [post-partum haemorrhage]”. In 2006, a WHO expert meeting recommended that “auxiliary nurse-midwives, community midwives, village midwives and health visitors...if they have been specially trained” can distribute misoprostol. However, in 2009, WHO’s Department of Making Pregnancy Safer issued a statement saying “WHO does not recommend distribution of misoprostol to community level health workers or women and their families for routine or emergency use”. The evidence base for establishing policy guidelines in the treatment of post-partum haemorrhage is disputed. What do we do when the type of randomised trial done by Widmer and colleagues is logistically, ethically, or fi nancially impossible to implement? A placebo cannot be used when we know misoprostol makes the uterus contract and no alternative therapy exists. It follows that historical or geographical controls are the only plausible source of evidence about home births without a skilled birth attendant. Are such See Articles page 1808