OBJECTIVE To review practice of massive primary postpartum haemorrhage management and develop a protocol. METHODS Cross-sectional study conducted at the Department of Obstetrics and Gynaecology at Aga Khan University Hospital, Karachi between January 1, 2003 and July 31, 2004. Women with primary postpartum haemorrhage and had blood loss > or = 1000ml were included in the study. Medical record files of these women were reviewed for maternal mortality and morbidities which included mode of delivery, possible cause of postpartum haemorrhage, supportive, medical and surgical interventions. RESULTS Approximately 3% (140/4881) of women had primary postpartum haemorrhage. 'Near miss' cases with blood loss > or = 1500ml was encountered in 14.37% (20/140) of these cases. Fifty-six percent (18/32) of the women who had massive postpartum haemorrhage delivered vaginally. Uterine-atony was found to be the most common cause, while care in High Dependency Unit (HDU) was required in 87.5% (28/32) of women. In very few cases balloon tamponade (2-cases) and compression sutures (2-cases) were used. Hysterectomy was performed in 4-cases and all of them encountered complications. Blood transfusions were required in 56% of women who had massive postpartum haemorrhage. CONCLUSION This study highlights the existence variable practices for the management of postpartum haemorrhage. Interventions to evaluate and control bleeding were relatively aggressive; newer and less invasive options were underutilized. Introduction of an evidence-based management model can potentially reduce the practice variability and improve the quality of care.