Research article

Misoprostol for treating postpartum haemorrhage: a randomized controlled trial [ISRCTN72263357]

G Justus Hofmeyr¹ , Sandra Ferreira¹ , V Cheryl Nikodem² , Lindeka Mangesi¹ , Mandisa Singata¹ , Zukiswa Jafta¹ , Babalwa Maholwana¹ , Zonke Mlokoti³ , Gijs Walraven⁴ and A Metin Gülmezoglu⁵

¹  Effective Care Research Unit, University of Witwatersrand and Fort Hare, and East London Hospital Complex, East London, South Africa

²  Department of Nursing, University of the Western Cape, Cape Town, South Africa

³  Tembisa Hospital Effective Care Research Unit, Tembisa, South Africa

⁴  Aga Khan Health Services, Aiglemont, France

⁵  HRP/RHR, World Health Organisation, Geneva, Switzerland

author email corresponding author email

BMC Pregnancy and Childbirth 2004, 4:16doi:10.1186/1471-2393-4-16

Published:	6 August 2004

Abstract

Background

Postpartum haemorrhage remains an important cause of maternal death despite treatment with conventional therapy. Uncontrolled studies and one randomised comparison with conventional oxytocics have reported dramatic effects with high-dose misoprostol, usually given rectally, for treatment of postpartum haemorrhage, but this has not been evaluated in a placebo-controlled trial.

Methods

The study was conducted at East London Hospital Complex, Tembisa and Chris Hani Baragwanath Hospitals, South Africa. Routine active management of the third stage of labour was practised. Women with more than usual postpartum bleeding thought to be related to inadequate uterine contraction were invited to participate, and to sign informed consent. All routine treatment was given from a special 'Postpartum Haemorrhage Trolley'. In addition, participants who consented were enrolled by drawing the next in a series of randomised treatment packs containing either misoprostol 5 × 200 μg or similar placebo, which were given 1 orally, 2 sublingually and 2 rectally.

Results

With misoprostol there was a trend to reduced blood loss ≥500 ml in 1 hour after enrolment measured in a flat plastic 'fracture bedpan', the primary outcome (6/117 vs 11/120, relative risk 0.56; 95% confidence interval 0.21 to 1.46). There was no difference in mean blood loss or haemoglobin level on day 1 after birth < 6 g/dl or blood transfusion. Side-effects were increased, namely shivering (63/116 vs 30/118; 2.14, 1.50 to 3.04) and pyrexia > 38.5°C (11/114 vs 2/118; 5.69, 1.29 to 25). In the misoprostol group 3 women underwent hysterectomy of whom 1 died, and there were 2 further maternal deaths.

Conclusions

Because of a lower than expected incidence of the primary outcome in the placebo group, the study was underpowered. We could not confirm the dramatic effect of misoprostol reported in several unblinded studies, but the results do not exclude a clinically important effect. Larger studies are needed to assess substantive outcomes and risks before misoprostol enters routine use.