Poster Session 4
Rachel L. Wiley, MD, MPH (she/her/hers)
MFM Fellow
University of California, San Diego
San Diego, California, United States
Ipsita Ghose, MD
Baylor College of Medicine
Houston, Texas, United States
Hector M. Mendez-Figueroa, MD
Associate Professor
McGovern Medical School at UTHealth
Houston, Texas, United States
Suneet P. Chauhan, MD
Director of MFM Research
Delaware Center of Maternal-Fetal Medicine at Christiana Care
Delaware, Delaware, United States
ACOG released a Committee Opinion stating quantitative blood loss (QBL) was more accurate than visually estimated blood loss (EBL), however the clinical impact of QBL remains uncertain.
Study Design:
This was a secondary analysis of a retrospective cohort study of all singletons delivered at > 20 weeks at a Level IV center during 24 months. This time period included an institutional transition from EBL to QBL, and patients with both methods of estimation recorded were included. Composite maternal adverse outcome included additional surgical techniques, transfusion of > 4 units, balloon tamponade, intensive care unit (ICU) admission, venous thromboembolism (VTE), hysterectomy or death. The diagnostic accuracy of CMAO for EBL and QBL were assessed by using the area under the curve (AUC) of receiver-operating characteristics (ROC) curves, and diagnostic test statistics were calculated and compared using overlapping confidence intervals.
Results:
Of 8,623 deliveries, 1,530 (18%) had both EBL and QBL recorded; the demographics are included in Table 1. QBL identified more postpartum hemorrhages (PPH; blood loss > 1000 mL) than EBL (9.0% vs. 6.0%; p< 0.01). CAMO occurred in 42 (3.0%) of patients, and Figure 1 shows the ROC curve for detecting CAMO with a nonsignificant difference in AUC of 0.91 for EBL and 0.86 for QBL (p=0.07). Using the threshold for PPH of > 1000 mL, EBL and QBL were equally sensitive, but EBL was more specific (95.8%; 95% CI 94.6% - 96.7% vs 92.6%; 95% CI 91.1% - 93.9%, Figure 1). While positive and negative predictive power and likelihood ratio did not differ significantly, negative predictive power and likelihood ratio performed better with EBL.
Conclusion:
While QBL and EBL are both predictive for maternal morbidity, estimated blood loss is less likely to meet the threshold for PPH, EBL is more specific and may predict adverse outcomes better. The benefits of QBL over EBL and clinical outcomes need to be evaluated in prospective trials.
