📄 obstetrics and gynecology

Comparing measured and calculated blood loss after a caesarean birth

This secondary analysis of the I'M WOMAN trial data reveals that while measured and calculated blood loss methods show weak correlation and calculated loss overestimates high-volume thresholds, measured blood loss demonstrates superior predictive ability for severe maternal outcomes, though calculated loss remains a viable objective alternative when direct measurement is not feasible.

Original authors: Mansukhani, R., Arribas, M., Bello, N., Chaudhri, R., Geer, A., Ker, K., Muganyizi, P., Prowse, D., Roberts, I.

Published 2026-07-18

📖 3 min read☕ Coffee break read

CC BY 4.0

Original authors: Mansukhani, R., Arribas, M., Bello, N., Chaudhri, R., Geer, A., Ker, K., Muganyizi, P., Prowse, D., Roberts, I.

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ⚕️ This is an AI-generated explanation of a preprint that has not been peer-reviewed. It is not medical advice. Do not make health decisions based on this content. Read full disclaimer

Imagine the human body as a bustling city, where blood is the vital traffic keeping everything running. When a major event happens, like a baby being born, sometimes the traffic gets a little chaotic, and too much of it spills out. This is called postpartum hemorrhage, and it's a serious emergency that can shut down the city if not fixed fast. For years, doctors have tried to guess how much traffic has spilled by just looking at the mess with their eyes. But here's the tricky part: human eyes are terrible at guessing volume. They tend to think the spill is much smaller than it really is, like looking at a flooded street and thinking it's just a puddle. Because of this, scientists have been trying to find better ways to measure the spill, either by weighing the wet towels and sucking up the liquid (measured loss) or by doing some math based on how much the "traffic density" (hemoglobin) drops in the blood before and after the event (calculated loss). Knowing the exact amount is crucial because if you don't know how bad the flood is, you might send the wrong amount of help, or send it too late.

This paper is like a massive detective story comparing two different ways to solve the mystery of how much blood is lost during a C-section. The researchers took data from over 10,000 women across Nigeria, Pakistan, and Tanzania to see which method tells the truth better. They compared the "weighing the towels" method against the "math formula" method.

Here is what they found: The two methods are like two friends who agree on the average size of a pizza but can't agree on the size of any single slice. On average, both methods said the blood loss was around 500 to 545 ml. However, when you look at individual women, the two methods barely agree with each other. It's as if one friend says, "You lost a small amount," and the other says, "You lost a huge amount," even though they are talking about the same person. The "math formula" method was much more variable; it often guessed that women lost way more blood than the "weighing" method did. In fact, the math method said three times as many women lost over 1,000 ml of blood compared to the weighing method.

The researchers also checked which method was better at predicting when a woman might get into serious trouble, like going into shock or needing a life-saving transfusion. The "weighing" method was a slightly better detective, spotting the danger signs a bit more accurately than the math method. But here is the twist: the math method has a big advantage. It doesn't need special scales or trained staff to weigh towels; it just needs a weight scale and a blood test, which are easier to find in many hospitals.

So, while the "weighing" method is the more accurate detective for predicting who might get sick, the "math" method is still a very useful tool, especially when you can't weigh the towels. The paper suggests that while the math method isn't perfect and often overestimates the loss, it remains a solid, objective way to measure blood loss for big studies when direct measurement isn't possible. The key takeaway is that we can't rely on our eyes to guess, and while we have two good tools, they tell slightly different stories, so doctors need to be careful about which one they trust in the heat of the moment.

Problem
Accurate measurement of blood loss following caesarean birth is critical for the early identification and treatment of postpartum haemorrhage (PPH), a leading cause of maternal mortality. While visual estimation is commonly used, it is known to underestimate actual blood loss, potentially delaying life-saving interventions. Although objective techniques are recommended by the World Health Organization, there is a need to compare different objective methods. Specifically, this study addresses the discrepancy between "measured blood loss" (direct quantification via weighing and suction) and "calculated blood loss" (derived from maternal weight and peripartum haemoglobin changes), particularly in the context of caesarean births in low-resource settings.

Methodology
This study is a secondary analysis of prospectively collected data from the I'M WOMAN trial, a double-blind randomised controlled trial investigating tranexamic acid for PPH prevention. The analysis included 10,393 women undergoing caesarean birth at 32 hospitals in Nigeria, Pakistan, and Tanzania.

Two methods were compared:

Measured Blood Loss (MBL): Obtained by weighing swabs, pads, and gauzes (1g = 1ml), quantifying suctioned blood (with amniotic fluid deducted), and measuring blood in calibrated drapes placed beneath the woman for 1–2 hours post-surgery, followed by visual estimation up to 24 hours.
Calculated Blood Loss (CBL): Derived using the Shook formula: $CBL = \text{Estimated Blood Volume} \times \frac{Hb_{prebirth} - Hb_{postpartum}}{Hb_{prebirth}}$ . Estimated blood volume was calculated as $0.085 \times \text{maternal weight (kg)}$ . Postpartum haemoglobin was measured at 24 hours (or discharge), and values were adjusted for red blood cell transfusions. Negative calculated values were replaced with zero.

Statistical analyses included median (IQR) comparisons, Spearman's correlation, Bland-Altman analysis for agreement, and Cohen's kappa for threshold agreement (300, 500, 1000, 1500 ml). The predictive ability of both methods for "death or near-miss" (defined as death from bleeding, massive transfusion, organ failure, or surgical intervention) and "haemodynamic compromise" (shock index >1.0) was assessed using Area Under the Curve (AUC) of Receiver Operating Characteristic (ROC) curves. Analyses were stratified by anaemia status.

Key Results

Correlation and Agreement: While median MBL (545 ml) and CBL (505 ml) were similar, the correlation between the two methods was weak (Spearman's rho = 0.28). Bland-Altman analysis revealed a small mean bias (-39 ml) but extremely wide limits of agreement (-1196 ml to 1118 ml), indicating poor individual-level agreement.
Threshold Classification: CBL classified significantly more women as having high blood loss compared to MBL. For losses $\ge$ 1000 ml, CBL identified 21% of women versus 7% for MBL. For losses $\ge$ 1500 ml, CBL identified 8% versus 2% for MBL. Cohen's kappa indicated weak to fair agreement across thresholds.
Anaemia Stratification: Women without anaemia showed a greater proportional haemoglobin drop for blood losses below 1000 ml compared to women with anaemia. Consequently, in women without anaemia, CBL was higher than MBL, whereas in women with anaemia, MBL was higher than CBL.
Predictive Ability: MBL demonstrated modestly superior predictive ability for adverse outcomes compared to CBL.
- Death or Near-Miss: MBL AUC was 0.87 vs. 0.76 for CBL ( $p<0.001$ ).
- Haemodynamic Compromise: MBL AUC was 0.66 vs. 0.62 for CBL ( $p=0.001$ ).
- Specificity improved with higher thresholds for both methods, but sensitivity decreased.

Significance and Claims
The paper concludes that while median measured and calculated blood losses are similar, the two methods are only weakly correlated and classify patients differently, particularly at higher thresholds. Calculated blood loss tends to overestimate the proportion of women with severe blood loss compared to direct measurement.

The authors claim that measured blood loss has a modest advantage in predicting death, near-miss, and haemodynamic compromise. However, they emphasize that calculated blood loss remains a valid objective measure suitable for use as a trial outcome when direct measurement is not feasible, as it relies on routinely collected data (weight and haemoglobin) and avoids errors associated with spillage, contamination, or the logistical challenges of weighing swabs in busy clinical settings. The study highlights the limitations of both methods, noting that measured loss can be contaminated by fluids and underestimated during heavy bleeding, while calculated loss assumes stable blood volume and is affected by haemodilution from intravenous fluids.

More like this