Approximately 260,000 women die each year from childbirth and pregnancy complications, mainly from postpartum hemorrhage (PPH), specifically in low-income regions like South Asia and sub-Saharan Africa, because of limited access to care. While blood loss criteria for PPH exist, they may not accurately reflect individual risk due to physiological variations. The same volume of blood loss may offer different risks for different women. To reduce preventable deaths, better diagnostic criteria and markers are important for early detection of high-risk women, specifically in resource-limited areas.
This study aims to evaluate the prognostic accuracy of clinical markers to predict maternal mortality and severe morbidity related to postpartum bleeding. It further examines whether a lower blood loss threshold, along with physiological markers like blood pressure, pulse rate, and shock index, can more effectively detect severe maternal outcomes and support early diagnosis and management of PPH.
Researchers conducted a systematic search and identified 33 potentially eligible datasets that involved 370,954 women who had given birth. Individual participant data were obtained from 18 datasets, which represent 314,547 women. Twelve datasets were included in the final analysis after quality assessment and exclusion of incomplete data, which represent 312,151 participants. 98.1% of women delivered vaginally, and the majority got uterotonic prophylaxis for PPH prevention. This study was assessed for bias and found to have a generally low risk of bias and good applicability.
Measured blood loss was determined at many thresholds (200 to 1000 mL) to assess predictive performance. The results showed that sensitivity decreased with increasing blood loss thresholds. Sensitivity was 89.8% at 200 mL, and it reduced to 47.9% at 1000 mL. Measured blood loss was achieved at a specificity of 81.4% and a sensitivity of 75.7% at a conventional threshold of 500 mL. The predefined sensitivity target of above 80% was achieved at a threshold of 300 mL, though with reduced specificity of 54.8%. In other clinical markers, diastolic and systolic blood pressure, shock index, and pulse rate showed variable performance, with diagnostic odds ratios (DORs) above 2, suggesting some discriminatory power.
None of these markers individually met the predefined target sensitivity and specificity criteria to justify their use as standalone predictors. Decision rule models, incorporating blood loss, were tested along with other markers. The model combining blood loss between 300 and 499 mL plus any abnormal clinical marker showed a balanced performance with a sensitivity of 87% and a specificity of 66.6% to 76.1%. It yields the highest DOR values (14.4 to 21.1). Subgroup analyses revealed that there is a better prognostic accuracy for vaginal births than for caesarean sections, with no major differences in country income levels or participant baseline risk.
These findings suggest that measured blood loss remains the most reliable marker of adverse maternal effects after childbirth. Decreasing the threshold to 300 mL from 500 mL can improve sensitivity to detect at-risk women. When combined with other abnormal hemodynamic markers, such as hypotension and tachycardia, the diagnostic accuracy increases without a significant loss of sensitivity. It also highlights the methods to measure blood loss, which significantly affect predictive accuracy, with weighted measurement to produce narrower and more reliable estimates. The large sample size and analytical method strengthen the validity of the results.
This meta-analysis supports lowering the PPH diagnostic threshold from 500 mL to 300 mL, particularly when combined with hemodynamic markers. This can improve the early detection of at-risk women. It further suggests that lowering PPH standards can improve maternal safety globally but highlights the necessity of conducting implementation studies to determine the efficacy and feasibility of these reduced thresholds in various healthcare settings, particularly in low-income regions.
Reference: Gallos I, Williams CR, Price MJ, et al. Prognostic accuracy of clinical markers of postpartum bleeding in predicting maternal mortality or severe morbidity: a WHO individual participant data meta-analysis. Lancet. 2025;0(0):0-0. doi:10.1016/S0140-6736(25)01639-3
