To estimate prevalence, indications, risk factors and complications associated with emergency postpartum hysterectomy and to compare different results in Teaching and District Hospitals we designed a retrospective analysis in 34 cases of postpartum hysterectomy performed from 1 January 2002 to 31 December 2006, in three different obstetrics and gynaecology units (one University Hospital and two Regional Hospitals). The prevalence of postpartum hysterectomies was 2.2 per 1,000 deliveries. District Hospitals revealed a significantly increased risk (OR: 2.28). Severe complications were few. No deaths were reported. Considering the totality of patients and cases of District Hospitals, placenta praevia was the main cause of hysterectomy (44.1% and 52.2%, respectively), but uterine atony had a major prevalence in the Teaching Hospital (45.4%). Operating time was significantly longer (p = 0.0001) and the blood loss was significantly more important (p = 0.041) in the Teaching Hospitals.