Poster Session 3
Category: Obstetric Quality and Safety
Poster Session 3
Arielle J. Higgs, MD (she/her/hers)
Fellow
University of Maryland
Baltimore, Maryland, United States
Sara Case, MSPH
University of Maryland
Baltimore, Maryland, United States
Jenifer Fahey, CNM, PhD
University of Maryland
Baltimore, Maryland, United States
Courtney Townsel, MD, MSc
Assistant Professor
University of Maryland School of Medicine
University of Maryland School of Medicine, Maryland, United States
Postpartum hemorrhage (PPH) remains a leading cause of maternal mortality worldwide. The American College of Obstetricians and Gynecologists defines PPH as blood loss ≥1000 mL, a threshold that commonly prompts clinical intervention, including consideration of transfusion. However, up to 40% of pregnant patients are anemic at term due to hemodilution and increased iron demands. These patients have reduced oxygen-carrying capacity and may require transfusion at lower blood loss volumes. Despite this, current protocols use universal blood loss thresholds, which underestimate transfusion need in vulnerable subgroups. This study evaluated diagnostic performance of pre-delivery hemoglobin in predicting peripartum transfusion and aimed to identify optimal hemoglobin thresholds for clinical decision-making.
Study Design:
This retrospective cohort study analyzed delivery data from 2019–2024 at a single tertiary center. After data cleaning, 8,379 patients with blood loss, pre-delivery hemoglobin, and transfusion data were analyzed. Patients were stratified into two groups: Group 1 (Hgb < 11 g/dL, n=3,504) and Group 2 (Hgb ≥11 g/dL, n=4,875). ROC curves were generated for each group to evaluate EBL as a transfusion predictor. Optimal thresholds were identified using the Youden Index, and performance metrics were calculated at EBL cutoffs (500–2000 mL).
Results:
Group 1 had an AUC of 0.814 for predicting transfusion, with an optimal blood loss threshold of 732 mL. Group 2 had an AUC of 0.877 and a threshold of 948 mL. Transfusion occurred in 14.6% of Group 1 and 3.8% of Group 2. At 1000 mL, Group 1 had lower sensitivity (0.55) but higher PPV (0.50) than Group 2 (sensitivity 0.73, PPV 0.23).
Conclusion:
Stratifying transfusion thresholds by hemoglobin improves sensitivity without sacrificing specificity. These findings support individualized transfusion protocols as fixed blood loss thresholds may inadequately reflect transfusion need in anemic patients. Hemoglobin-informed thresholds may optimize clinical response to peripartum blood loss before conventional PPH criteria are met.