Objective To evaluate the ability of estimated blood loss (EBL) and quantitative blood loss (qBL) to predict need for blood transfusion in women with postpartum hemorrhage (PPH).
Study Design This is a retrospective chart review that identified women with PPH (>1,000 mL for vaginal or cesarean delivery) between September 2014 and August 2015, reported by EBL (n = 92), and October 2015 and September 2016, reported by qBL (n = 374). The primary metric was the area under the receiver-operating characteristic curve for blood transfusion.
Results The rate of PPH by EBL and qBL was 2.8 and 10.8%, respectively (p < 0.01). The rate of transfusion for women meeting criteria for PPH by EBL and QBL were 2% (66/3,307) and 2.7% (93/3,453), respectively (p = 0.06). Postpartum transfusion was predicted by an EBL of 1,450 mL with AUC 0.826 and qBL 1,519 mL with AUC 0.764, for all modes of delivery. Postpartum vital signs and change in pre- and postdelivery hematocrit were poor predictors for transfusion.
Conclusion The rates of PPH increased with the implementation of qBL. Overall, qBL did not perform better than EBL in predicting the need for blood transfusion.