A population-based review (1988-2000) of 142,634 deliveries found an incidence of emergency obstetric hysterectomy of 0.53 per 1000 deliveries. The relative risk for caesarean versus vaginal delivery was 18.32 (10.26, 32.71) P<0.001. Previous caesarean delivery occurred in 47.4% and 22.4% were primiparous. Causes and morbidity were ascertained in a hospital-based study (1980-2001) of 64 emergency hysterectomies in 110,537 deliveries. The main indications for hysterectomy were abnormal placentation (50%) and atonic postpartum haemorrhage (32.8%). Blood transfusion was needed in 84.4% and intensive care in 26.6%. There were no maternal deaths. Potential alternatives to hysterectomy for massive obstetric haemorrhage are discussed.