Investigation of an increase in postpartum haemorrhage in Canada

OBJECTIVE: To investigate the cause of a recent increase in hysterectomies for postpartum haemorrhage in Canada. DESIGN: Retrospective cohort study. SETTING: Canada between 1991 and 2004. POPULATION: All hospital deliveries in Canada as documented in the database of the Canadian Institute for Health Information (excluding incomplete data from Quebec, Manitoba and Nova Scotia). METHODS: Deliveries with postpartum haemorrhage by subtype were identified using International Classification of Diseases codes, while hysterectomies were identified using procedure codes. Changes in determinants of postpartum haemorrhage (all postpartum haemorrhage and that requiring hysterectomy) were examined, and crude and adjusted period changes were assessed using logistic models. MAIN OUTCOME MEASURES: Postpartum haemorrhage, postpartum haemorrhage with hysterectomy, postpartum haemorrhage with blood transfusion and postpartum haemorrhage by subtype. RESULTS: Rates of postpartum haemorrhage increased from 4.1% in 1991 to 5.1% in 2004 (23% increase, 95% CI 20-26%), while rates of postpartum haemorrhage with hysterectomy increased from 24.0 in 1991 to 41.7 per 100,000 deliveries in 2004 (73% increase, 95% CI 27-137%). These increases were because of an increase in atonic postpartum haemorrhage, from 29.4 per 1000 deliveries in 1991 to 39.5 per 1000 deliveries in 2004 (34% increase, 95% CI 31-38%). Adjustment for temporal changes in risk factors did not explain the increase in atonic postpartum haemorrhage but attenuated the increase in atonic postpartum haemorrhage with hysterectomy. CONCLUSIONS: There has been a recent, unexplained increase in the frequency, and possibly the severity, of atonic postpartum haemorrhage in Canada.