Performance of quantitative and graduated estimated blood loss in detection of postpartum hemorrhage after spontaneous vaginal delivery.

Objective: To compare the performance of quantitative blood loss (QBL) to graduated estimated blood loss (EBL) for prediction of clinical outcomes in spontaneous vaginal deliveries (SVDs). Methods: Prospective observational study including SVDs with a recorded QBL at a single institution. QBL was calculated and recorded in the electronic medical record by the nurse following delivery with the providers pragmatically blinded to these calculations. Providers used clinical indices and a graduated EBL, assigned prior to QBL calculation, to diagnose postpartum hemorrhage (PPH) with a specific postpartum acuity-based observation process for patients with PPH. PPH was defined as blood loss >500 cc. EBL and QBL were directly compared for the purposes of this investigation. Performance characteristics were calculated for EBL and QBL with regard to detection of transfusion and decrease in hematocrit greater than 10% (ΔHct ≥ 10) from the day of delivery to postpartum day one. Patients with discrepancies between QBL and EBL were examined with odds ratios of transfusion and ΔHct ≥ 10 calculated. Results: There were a total of 8107 SVDs. Rates of PPH were higher when defined by QBL (17% vs. 9%), with a 2.5% rate of transfusion. A discrepancy >500 cc between QBL and EBL was found in 226 (3%) patients. PPH as defined by EBL demonstrated lower sensitivity but higher specificity, PPV, and NPV for transfusion and higher specificity and PPV for ΔHct ≥ 10. When examining deliveries with EBL ≤ 500 cc but QBL ≥ 500 cc, odds of transfusion (odds ratio [OR], 1.41; 95% confidence interval [CI], 1.02-1.93) and ΔHct ≥ 10 (OR, 2.94; 95% CI, 2.39-3.59) were increased. Conversely, this was true in cases of EBL ≥ 500 cc with a QBL ≤ 500 cc, with an odds of transfusion (OR, 6.38; 95% CI, 4.43-9.01) and ΔHct ≥ 10 of 3.41 (95% CI, 2.41-4.47). For every 2700 deliveries in which QBL was used to guide care, one additional transfusion would be identified versus using EBL alone. Conclusion: Graduated EBL demonstrates higher specificity for transfusion and ΔHct ≥ 10 when compared to QBL. In an environment with a standardized graduated EBL and an acuity-based postpartum observation process, QBL does not offer additional benefits after SVD. In places with limited resources, graduated EBL is a good alternative to QBL.