Key points are not available for this paper at this time.
INTRODUCTION: Postpartum hemorrhage is the leading cause of maternal mortality worldwide. Quantitative blood loss calculators have been developed to improve recognition of hemorrhage; however, there is a paucity of research comparing their accuracy. The objective of this study is to compare quantitative blood loss and estimated blood loss values during vaginal deliveries with calculated blood loss determined by maternal hematocrit before and after delivery. METHODS: This IRB-approved retrospective cohort study included all term, spontaneous vaginal deliveries of singleton infants to patients 18 years or older who delivered at a single community-based tertiary center between August 2019 and December 2020. Estimated blood loss (EBL) or quantitative blood loss (QBL) were obtained from delivery notes. Quantitative blood loss was obtained using the Triton Gravimetric Calculator. Data were abstracted from electronic medical records and de-identified. Calculated estimated blood loss (cEBL) was determined using the formula determined by Stafford et al. in 2008. The associations between EBL and cEBL and between QBL and cEBL were evaluated by using linear regression. RESULTS: Of 1,667 deliveries that met inclusion criteria, 670 had a QBL (40.2%) and 997 had an EBL (60.9%). The association between cEBL and QBL using simple linear regression was a poor fit with an R-square of 0.206. Estimated blood loss and cEBL also had a poor association using linear regression with an R-square of 0.0941. By using R-square, QBL can predict better than EBL; however, both are poor predictors of blood loss calculated from pre-delivery and post-delivery hematocrit. CONCLUSION: This study demonstrates that both QBL and EBL have poor correlation with the decrease in hematocrit after a vaginal delivery. Both QBL and EBL missed a significant blood loss in one out of three deliveries. A prospective, blinded, cohort study is currently in process to directly compare QBL and EBL at our institution.
Winograd et al. (Wed,) studied this question.