There is two formulas admitted in the medical literature: The Mercurial formula, designed to calculate the blood loss in ml of pure hb (100% ht) ,ad the Gross formula ubased on the assumption of a linear decrease of hb during hemorrhage

Thanks very much - very helpful.

I have a somewhat more simplistic approach that I have found useful in the ICU. It is not nearly so sophisticated (nor does likely have as high an r2) as the nicely documented formulas provided in the papers.

The first assumption is that you start with a hematocrit of 30% (hemoglobin of 10 mg/dL). Each unit of blood loss drops the hematocrit by 3 percent points (hemoglobin by 1 mg/dL). Conversely stated if the hematocrit drops by 6% the patient has lost 2 units of blood.

For each unit of blood transfused, you should expect to see the hematocrit increased by 3 percentage points and hemoglobin increased by 1 mg/dL.

Therefore, a patient who enters the operating room with a hematocrit of 30% (hemoglobin 10 mg/dL) and comes out of the operating room with the same hematocrit of 30% but has been transfused 10 units has an estimated blood loss of one blood volume or approximately 5 L.

This paradigm works on the following relatively crude assumptions/calculations.

   1) each unit of packed red blood cells is roughly 250 mL with a hematocrit of 60%

   2) from that assumption if that unit is mixed 1:1:1 ratio with plasma and platelets the result is a 500 mL bag with a hematocrit of 30% (10 mg/dL Hbg)

3) since the expected human blood volume is 70 mL/kg, a 70 kg man should have a blood volume of roughly 5 L. If each unit is roughly equivalent to 500 mL  of whole blood then 10×500 mL is 5 L (or one human blood volume.)

4) A full human blood volume of 5 L then equates to 10 units of blood when given in the appropriate 1:1:1 ratio with plasma and platelets.

This is not formula that has precision or accuracy to be used in research but for the practicing academic physician (surgeon) who needs to do an approximation in the middle of the night, it works.

I confess my sin of being a surgeon so that when the experts in the light of day look at this and point out all its flaws they can just write it off to another crazy surgeon. But it has worked for me for 30 years.

I agree with John Owings, it¨s a simple calculation, considering the intial hematocrit or best hemoglobin concentration and the final one, you may conclude aproximately the blood loss.

I still do not understand how such an interesting question (and especially its answer) does not have sufficiently accurate proposals, in the clinical and surgical setting, to estimate the bleeding, nor by clinicians or the monitoring companies. At least, if we want to estimate intraoperative blood loss, where intravascular volume may not be constant.

Article Correlation of transfusion volume to change in hematocrit