Accordance between weight-based and calculated blood loss for obstetric hemorrhage: a pre-study to a randomized controlled trial.

Abstract

Objective: Postpartum hemorrhage (PPH) is the leading cause of maternal death. Identifying excessive bleeding is crucial, but there is no gold standard for its measurement. We aimed to assess accordance between two formulas for calculated blood loss (CBL) and weight-based blood loss (WBBL) to determine if CBL has a place in either the clinical or research obstetric context.

Method: This was a pre-posttest study nested in a pilot study. Women ≥ 18years and ≥ 36 gestational weeks with uncomplicated pregnancies and planned vaginal birth were included at a tertiary hospital in Stockholm, Sweden from December 2022 to February 2023. Hematocrit (Erythrocyte Volume Fraction, EVF) and hemoglobin (Hb) were taken at admission and 24 h postpartum. Blood loss was calculated in three ways: weighing blood collected in drapes and blood-soaked material; Gross's formula based on pre-postpartum EVF; and pre-postpartum Hb difference, assuming a 10 mg/L difference equivalent to 500 ml. We compared median blood loss between these methods using Related-samples Wilcoxon Signed-Rank Test, correlations using Spearman's Rho, and prediction of PPH (≥500 ml) using McNemar's test.

Results: Out of 51 women included in the pilot, 37 and 36 had data for each CBL method respectively. Median blood loss was 350 ml for WBBL, 649 ml for Gross's formula (p = 0.008) and 750 ml for pre-postpartum Hb difference (p < 0.001). For the diagnosis PPH, WBBL and Gross's formula were concordant in 61% of cases (p = 0.002), compared to 43% for WBBL and pre-postpartum Hb difference (p = 0.001). There was a moderate positive correlation between WBBL and both CBL methods (p = 0.001). Excluding women who received intravenous fluids did not improve agreement between WBBL and CBL methods.

Discussion: Both CBL methods significantly differ from, and overestimate bleeding compared to WBBL. There is a correlation between WBBL and CBL, but the estimates differ to the extent that CBL methods do not seem relevant for clinical obstetric practice or research.