Measuring post-partum haemorrhage in low-resource settings: The diagnostic validity of weighed blood loss versus quantitative changes in hemoglobin

Abstract

Background: Accurate estimation of blood loss is central to prompt diagnosis and management of postpartum hemorrhage (PPH), which remains a leading cause of maternal mortality in lowresource countries. In such settings, blood loss is often estimated visually and subjectively by attending health workers, due to inconsistent availability of laboratory infrastructure. We evaluated the diagnostic accuracy of weighed blood loss (WBL) versus changes in peri-partum hemoglobin to detect PPH. Methods: Data from this analysis were collected as part of a randomized controlled trial comparing oxytocin with misoprostol for PPH (NCT01866241). Blood samples for complete blood count were drawn on admission and again prior to hospital discharge or before blood transfusion. During delivery, women were placed on drapes and had pre-weighed sanitary towels placed around their perineum. Blood was then drained into a calibrated container and the sanitary towels were added to estimate WBL, where each gram of blood was estimated as a milliliter. Sensitivity, specificity, negative and positive predictive values (PPVs) were calculated at various blood volume loss and time combinations, and we fit receiver-operator curves using blood loss at 1, 2, and 24 hours compared to a reference standard of haemoglobin decrease of >10%. Results: A total of 1,140 women were enrolled in the study, of whom 258 (22.6%) developed PPH, defined as a haemoglobin drop >10%, and 262 (23.0%) had WBL ≥500mL. WBL generally had a poor sensitivity for detection of PPH (<75% for most volume-time combinations). In contrast, the specificity of WBL was high with blood loss ≤ 500mL at 1h and ≤750mL at any time points excluding PPH in over 97% of women. As such, WBL has a high PPV (≤85%) in high prevalence settings when WBL exceeds 750mL. Conclusion: WBL has poor sensitivity but high specificity compared to laboratory-based methods of PPH diagnosis. These characteristics correspond to a high PPV in areas with high PPH prevalence. Although WBL is not useful for excluding PPH, this low-cost, simple and reproducible method is promising as a reasonable method to identify significant PPH in such settings where quantifiable red cell indices are unavailable.