To provide guideline for the clinical management of postpartum hemorrhage, and to build up a scoring system for the prediction of severe postartum hemorrhage.

This study reviewed 375 pregnant women with severe postpartum hemorrhage, who were treated on the Third Affiliated Hospital of Guangzhou Medical University from January 2000 to June 2008. Then we analyzed the relationship between scores and the maternal outcomes, which included organ dysfunction, hysterectomy and death, and established the prediction scoring system for postpartum hemorrhage. The data were analyzed using χ² test, Fisher exact probability test, linear association and logistic regression.

log eY / (1-Y) = -7.274 + 0.858 × prenatal score + 0.821 × postpartum score was the formula of prediction for sever postapartume hemorrhage. The score was 4 which suggested the incresing risk of organ dysfunction (χ²=13.23, P<0.01), and 12 scores suggested the incresing risk of three or more organs dysfunction (χ²=54.21, P<0.01). Over 8 scores suggested the risk of hysterectomy was increased significantly(χ²=20.65, P<0.01). There were significantly difference of serve postpartum hemorrhage between score 6 group and score 1-5 group(χ²=10.85, P<0.01). And sever postpartum hemorrhage also had significant difference between score 8 group and 6-7 group(χ²=4.68, P<0.05), score 12 group and score 8-11 group(χ²=10.53, P<0.01).

The prediction scoring system can be used to predict effectively the risk of severe postpartum hemorrhage and evaluate the serious degree of situation. In the further, we need a research on a large scale.