In search of the lost blood

One consequence of our evolving understanding of the safety and efficacy of allogeneic blood transfusion is the appreciation of the patient’s own blood as a highly valuable and unique resource that should not be wasted and discarded in vain. Losing one’s own blood will expose patients to the perils of anaemia, and replacing it with donor blood opens the door to a host of unfavourable outcomes¹. Accordingly, the role of clinicians is evolving as a proactive steward of patients’ limited and valuable physiological resources in order to achieve best clinical outcomes (as opposed to being mere reactive prescribers of treatments). This is reflected in the concept of Patient Blood Management (PBM), defined as “the timely application of evidence-based medical and surgical concepts designed to maintain haemoglobin concentration, optimise haemostasis and minimise blood loss in an effort to improve patient outcome”².

While reducing blood loss is a key component of PBM by helping to eliminate the need for transfusion in many patients, the volume and rate of blood loss remain important determinants in making transfusion decisions, as reflected in numerous guidelines³. Increased blood loss also serves as an important warning sign for complications such as postpartum haemorrhage, which may require immediate and drastic interventions. Finally, blood loss serves as an important direct end point in studies on various haemostatic measures and other interventions that might affect coagulation or bleeding. Needless to say, be it a study end point, a decision criterion, a warning sign, or a therapeutic target, accurate and objective assessment of surgical blood loss is currently an unmet need. In clinical practice, the most common way of assessing this important parameter is through visual estimation by the clinicians, which is known to be inaccurate and unreliable^4,5.

A distinction should be made between the assessment of blood loss (sudden or cumulative) as part of the clinical management of a patient (which is often needed in real time or within a short time in order in order to be useful in clinical decision making) and quantifying blood loss for research purposes (which often does not need to be immediately available and can be done “after-the-fact”). The study by Jaramillo et al. published in this issue of Blood Transfusion addresses the latter type of assessment⁶.

Drawing on the data from one hundred patients undergoing laparoscopic urological procedures, they used the difference between the preoperative haemoglobin level and nadir postoperative haemoglobin level to calculate the haemoglobin mass loss. They then compared this value with reference measures (direct measurements of blood loss volume and corresponding haemoglobin mass loss) and reported the agreement⁶. The idea is quite simple and has been considered before⁷. The haemoglobin mass loss can be calculated by multiplying the total blood volume by the difference in haemoglobin concentrations before and after the surgery. This is done with certain important assumptions: (i) haemoglobin distribution throughout the circulation should have reached equilibrium at the time of sampling; (ii) the haemoglobin measurement method should be reasonably accurate; (iii) the total blood volume should be relatively stable and estimated with relative accuracy; and (iv) there should have been no addition (e.g., blood cell transfusion). The final number should be further adjusted for the units of measurement (e.g., divided by 100 to yield haemoglobin mass loss in grams, if blood volume was provided in mL and haemoglobin values were in g/dL).

Not surprisingly, the main issue with this approach lies in the validity of the assumptions and measurements behind this formula. Various methods to measure haemoglobin concentration come with limitations in accuracy, and there is often a trade-off between the convenience of faster availability at one end and assay performance at the other. According to the standards set by the College of American Pathologists (CAP) and the Clinical Laboratory Improvement Amendments (CLIA), haemoglobin analysers used in the field are expected to stay within a somewhat arbitrary ±7% threshold of variation from the reference values⁸. The marketing materials for the analyser used in this study (HemoCue® Hb-201; HemoCue America, Brea, CA, USA) claim an accuracy of ±1.5% compared with the reference method. Reported variations in the studies range from −1.5 to+1.6 g/dL in venous blood (approximately ±10%, assuming an average haemoglobin of 15 g/dL), but most studies stayed within the ±7% variation range⁸. However, one should remember that, in a theoretical scenario in which one of the measurements is over-estimated while the other one is under-estimated versus the actual values, the variation in the difference of the preoperative and postoperative haemoglobin values from the difference of the reference values can inflate to just below 14% while still meeting the individual 7% CAP/CLIA criterion. Applying a worst-case scenario of ±7% variations in the mean pre- and postoperative values reported in this study (preoperative haemoglobin 144 g/L and postoperative haemoglobin 118 g/L, which yield a calculated estimated mass loss of 141.6 g considering the mean estimated blood volume as 5.448 L), the calculated haemoglobin mass loss can go as low as 42 g or as high as 241 g. While this is an extreme hypothetical case, it goes to show how relatively small and generally “accepted” errors in measurement of haemoglobin levels can have far reaching consequences in the calculated mass loss number.

A similar issue affects the estimated blood volume, which is the other essential component of this formula. Compared with direct methods of measurement, which are usually cumbersome and involve infusing radioactive labelled blood cells or indicators such as dyes, formulae used to estimate blood volume based on patient’s gender, body weight, height, and other demographics are convenient and widely used. Nonetheless, these formulae (including the one used in this study) are prone to frequent under- and overestimation⁹. Lastly, blood and plasma volumes in any given patients are not static parameters but undergo fluctuations under physiological conditions, as well as during procedures and in response to various factors such as hypoxia¹⁰. When we are talking about a patient’s “blood volume”, we are referring to a parameter that is difficult to measure, and even if it is measured with acceptable accuracy at a given time, it might, and frequently does, change moments later.

Challenges of dealing with an imperfect world should not prevent us from using this and other similar models in research and clinical practice. Einstein’s theory of general relativity is far more accurate than Newtonian mechanics, yet the latter is much simpler and capable of explaining many phenomena we encounter every day with acceptable accuracy, making it anything but obsolete. Likewise, the approach described by Jaramillo et al. has tremendous value in research given its simplicity and “acceptable” accuracy, as portrayed in their study⁶.

As we refer to blood loss volume, one should constantly ask what “blood” we are talking about: is this the “blood” in the circulation of patients just before the surgery without any loss or fluid replacement, at some point during surgery when it is partially diluted, suctioned and collected in a canister with additional fluids added, or after the surgery when all losses and additions have taken place and a new equilibrium is reached? Using haemoglobin mass, as Jaramillo et al. demonstrated, has the advantage of eliminating the effect of dilution and volume changes, making it much easier to trace and compare over time. On the other hand, we should remember that haemoglobin mass does not equate with red blood cell (RBC) mass (as haemoglobin is just one component of RBCs), and blood is not just the RBCs either (as it also includes platelets, plasma and its many factors, and more). Hence, tracing just the haemoglobin mass might not provide a comprehensive picture.

Perhaps the most important lesson to learn here is the advantages of using “mass” instead of “concentration” when referring to haemoglobin measurements. Currently, when haemoglobin is used in clinical practice (e.g., to define anaemia, make a transfusion decision, or set a treatment goal), it is expressed as a concentration (mg/dL or g/L). As can be seen from the study by Jaramillo et al.⁶, a concentration can have little clinical value without considering the total volume of the fluid that has that concentration. A patient can lose significant amounts of blood yet the haemoglobin concentration might not reflect that loss immediately. Haemoglobin concentration (or haematocrit) can remain deceivingly elevated in hypovolemic patients for several hours after the surgery/blood loss¹¹. More importantly, a similar amount of blood loss can have a very different impact on two different patients with the same baseline haemoglobin concentration of 13 g/dL, if one had substantially less total blood volume (as is often the case in women compared to men).

In our view, this study can serve as a reminder to us all to try to rethink the way we quantify and incorporate important measures related to anaemia and blood loss in our practice. Reporting values as concentration might be appropriate, and even physiologically sensible for other parameters such as electrolytes, as they distribute freely across extracellular space and their function is determined by their molar concentration. However, when it comes to blood and its many vital roles in the body, relying on haemoglobin concentration or haematocrit will leave us with a very limited and obscured view of the patient’s blood oxygen-carrying capacity and the real need for intervention and treatment of anaemia. This is particularly important given the common practice of over-transfusion in the face of documented risks and questionable benefits¹². We believe anaemia should be redefined, not as falling below an arbitrary haemoglobin concentration derived from epidemiological studies¹³, but as a disease marked by reduced capacity of blood to maintain its vital physiological functions.