To Transfuse or Not to Transfuse: That is the Question

Transfusion of blood is a common practice in both emergency departments, hospital wards, ICUs, and peri-operative settings worldwide; in the ED, we transfuse different kinds of patients all the time including but not limited to the trauma patient and GI/vaginal bleeder in hemorrhagic shock, and the undifferentiated symptomatic and sometimes asymptomatic anemic patient, among others. Let’s talk about the latter groups as transfusing patients who are actively bleeding in front of you is a bit more of a clear cut indication, although a recent trial on upper GI bleeders even challenges the limits of that indication1. Nearly 24 million blood components are transfused each year2. It has been noted that transfusion, along with cardiac stents, labor induction, ear tubes, and antibiotics are identified as one of five overused medical interventions by The American Medical Association3. And this stuff is certainly not cheap, not exactly benign, and is relatively hard to come by. So let’s talk about blood. Why? Because it’s cool and it’s important for us as clinicians taking care of sick patients to really know the risks and benefits of performing this particular intervention. Ultimately this is all about truly informed consent, and I find myself frequently wondering if I should actually be ordering that next unit of blood, what risks I am subjecting my patient to, and whether or not my patient is actually benefiting from that transfusion. The next few questions helped me understand this interesting yet complex topic a bit more.

1) First things first: what are we trying to accomplish by transfusing blood into a patient?

This really boils down to oxygen delivery to tissues. Marino (author of The ICU Book4) explains it beautifully and with cool graphs and figures: the fundamental goal of the organism, and in this case, our patients, is to deliver oxygen to individual cells2,4,5,. That’s why breathing is essentially an automatic thing; we don’t really have to make sure we take that next breath to oxygenate pulmonary capillary blood, there’s just not that much active thinking involved in this process. RBCs, and more specifically hemoglobin, are loaded up with oxygen molecules from the environment and this is how we deliver oxygen to cells and tissues, assuming we have a beating heart and a non-leaky circulatory system of course. Oxygen can then serve as the final electron acceptor in the electron transport chain of mitochondria and yield even more ATP for cellular processes that require energy5. Oxygen is either bound to hemoglobin or directly dissolved in blood and so we can calculate the oxygen content of blood by looking at these variables in the following equations4:

1) Oxygen Content of Blood=(1.34 x [Hb] x O2sat) + PaO2 x 0.003.

Simply including the patient’s cardiac output and we get:

2) Oxygen Delivery=CO x Oxygen Content of Blood.

3) Oxygen Uptake= CO x (Oxygen content of arterial blood-Oxygen content of venous blood).

4) Oxygen extraction ratio= Oxygen Uptake/Oxygen Delivery

Now what does a low hemoglobin/hematocrit (i.e. anemia) do? The body compensates by increasing the cardiac output and extracting more oxygen from the capillaries to the tissues, among other physiological responses, some of which include increased renal production of EPO, increased minute ventilation, and decreased systemic vascular resistance2,4,5. Increased oxygen extraction can be detected as a decrease in the mixed venous oxygen saturation4,5. The compensation taking place in the body is what at times leads to the clinical signs and symptoms of an anemia e.g. tachycardia, dyspnea, fatigue which we clinically detect through history and physical examination of the patient2,4,5. But should we intervene (i.e. transfuse blood) on the body’s evolutionarily-designed compensatory mechanisms as they run their natural course to correcting things back to homeostatic ranges; is this truly “pathology” when the body is responding as indeed it was ever so choreographed to do, or is this an indication that things are working just as they were intended to in the first place? A bit philosophical but discuss nonetheless.

The thing with transfusion of PRBCs is that we transfuse mainly based on the hemoglobin/hematocrit without knowing actually if we have reached the limits of the patient’s compensation when actual tissue injury occurs; i.e. we haven’t actually increased the patient’s tissue oxygen delivery at all by transfusing them, but rather have added more volume (each unit being approximately 350 mL, or between 300-400mL4,6) and subjected them to the risks (see next question) that come with transfusion of each unit of blood (See Paul L. Marino’s chapter on blood transfusion for more on this fascinating pathophysiology in The ICU Book4 and a great EB Medicine review6). Now anemia in one patient is certainly not the same anemia in another; e.g. the acute in onset hypovolemic anemia of the hemorrhaging GI bleeder or trauma patient deserves a different management strategy compared to the chronically anemic patient7.


2) What are the risks that are associated with blood transfusions?

Here are causes of badness following transfusion and why we should really be sure we want to go ahead and transfuse our patient.

The Non-Infectious Complications:

Transfusion-Related Acute Lung Injury (TRALI)-acute in onset hypoxemia with bilateral infiltrates on CXR, and no other explanation for this. Fun fact: first described in the 50s, this syndrome may be associated with the use of blood products derived from multiparous women as the preferential use of male or never been pregnant female-derived blood products led to a decrease in TRALI’s overall incidence6. TRALI is associated with 30-45% of reported transfusion-related mortality and is the most common cause of mortality and morbidity related to blood transfusions. Incidence is approximately 1 in 10 to 100 thousand transfusions8,9.

Transfusion-Associated Circulatory Overload (TACO)- exactly what it sounds like, too much volume leading to legitimate cardiogenic pulmonary edema as opposed to the non-cardiogenic pulmonary edema of TRALI6. Incidence nears 1% of transfusions with 13-34% of transfusion-related deaths being TACO-related8,9. Doesn’t sound like much but think about how many times you ordered units to be transfused last year, last five years, last decade; it’s all a numbers game!

Acute Hemolytic Transfusion Reaction (HTR)-most commonly due to ABO blood antigen incompatibility. The recipient essentially mounts an acute immune response to the donor RBCs leading to acute hemolysis with massive release of nephrotoxic hemoglobin, with subsequent acute renal failure and its associated consequences, and DIC6; yikes, this can quickly spiral out of control, not good! Incidence of 1 in 10 to 50 thousand transfusions8,9. Over the past 5 years both ABO and non-ABO HTRs have been associated with 22% of transfusion-related mortality8.

Delayed Hemolytic Transfusion Reaction-except for sickle cell patients this is usually only noted by a drop in hematocrit occurring up to 2 weeks following initial transfusion6.

Febrile Non-Hemolytic Transfusion Reaction-essentially just a fever, but don’t forget your differential diagnosis in this setting! Occurs in about 1 in 100 to 500 transfusions7,9.

Allergic Reactions-Once again, we are endowed with an immune system which at times seems to have a mind of its own; pesky IgE molecules! Respect the immune system response, because there is a very fine line between defense against invasion and all out chaos and coup d’etat! Allergic reactions can range from simple pruritis to full blown anaphylaxis with all associated complications. Anaphylaxis occurs in about 1 in 20 to 50 thousand transfusions7.

Transfusion-Associated Graft-Versus-Host Disease-this occurs when donor T cells mount an immune response against recipient tissues and organs; pick an organ, any organ, it is getting messed up bad. This rare entity which can be prevented by blood product irradiation carries a poor prognosis with a high mortality rate7.

The Infectious Complications:

Transmission of the following viruses:

HIV at a rate of 1 in 2.3 million, Hepatitis B virus at a rate of 1 in 350 thousand, Hepatitis C at a rate of 1 in 1.8 million, HTLV 1 and 2 at a rate of 1 in 2 million7.

Bacterial contamination with Yersinia being the most common occurs at a rate of 1 in 14 to 28 thousand7.

Other infectious risks include transmission of West Nile virus, Trypanosoma cruzi (the agent that causes Chagas Disease), CMV, syphilis, Parvovirus B19, Dengue fever, malaria, Plasmodium, and several others3.

With all of that said and done, and to put things into perspective, in 2011 in the USA there were a total of 30 deaths reported to the FDA that were attributed to blood transfusions while in 2012 and 2013 there were 388.


3) Now to the crux of the matter: to transfuse or not to transfuse

Since the 1999 publication of the Transfusion Requirements in Critical Care (TRICC)10 trial the literature has been piling up with papers supporting the use of a restrictive versus liberal approach to transfusing blood; prior to TRICC, ICU patients were transfused once the hemoglobin dropped below 10 g/dL. As you can imagine, many more units of blood were transfused with all associated costs and risks. The following are just a few studies in the post-TRICC era of transfusion practice I found to be good reads in support of the restrictive strategy:

Carson et al (2015)11 conducted a secondary analysis (looking at long-term mortality of the original FOCUS trial participants) of the Functional Outcomes in Cardiovascular Patients Undergoing Surgical Hip Fracture Repair (FOCUS) trial which looked at patients with cardiovascular disease or at risk for cardiovascular disease undergoing hip surgery and randomized them into a restrictive versus a liberal transfusion group12. They then followed them with a median 3.1 years follow up and looked for long-term mortality differences11,12. In FOCUS they randomly assigned 2016 elderly patients into 1009 patients allocated to a restrictive strategy (threshold Hb of 8 g/dL or with symptoms of anemia) versus a group of 1007 allocated to a liberal transfusion strategy (threshold Hb of 10 g/dL), both groups with a similar distribution of baseline characteristics (Older than 50 years old with a majority over 75 years old, with cardiovascular disease or cardiac risk factors, i.e. sounds like pretty sick patients)12. They found no clinically significant difference in overall long-term mortality between patients randomized to the liberal versus the restrictive blood transfusion groups; 43.2% died in the liberal transfusion group versus 40.8% in the restrictive transfusion group, with a hazard ratio of 1.09, 95% CI 0.95-1.25, P=0.2111. Of note, although both study arms were transfused, the liberal transfusion group was transfused approximately three-times more units11,12. In the original FOCUS trial there was no statistically significant difference in rates of death at 30 day follow up: 5.2% in the liberal transfusion group versus 4.3% in the restrictive transfusion group (absolute risk difference of 0.9%, 99% CI -1.5 to 3.4)12.

In the Transfusion Requirements after Cardiac Surgery (TRACS) trial, a prospective, randomized, controlled non-inferiority trial, Hajjar et al studied a restrictive (Hct<24%) versus liberal (Hct<30%) transfusion strategy in post-op cardiac surgery patients with anemia and detected no difference in the composite primary endpoint of 30-day all cause mortality and morbidity; 10% (95% CI 6-13%) in the liberal versus 11% (95% CI 7-15%) in the restrictive transfusion arms (between group difference of 1%, 95% CI -6 to 4%, p=0.85)13. 30 day mortality rates were 5% in the liberal (95% CI 2-7%) versus 6% in the restrictive group (95% CI 3-9%, P=0.93). An interesting finding from this study is that the number of transfused units of PRBCs was a statistically significant independent risk factor for cardiac, respiratory, renal and infectious complications13. In fact transfusing at least 5 units of blood increased mortality13.

Shander et al14 looked at patients with low post-op hemoglobin levels who refused to receive blood products and their mortality (n=293 with a majority of 288 Jehovah’s Witnesses, undergoing various kinds of surgical procedures). I found this to be very interesting as a study that shows what happens when we don’t transfuse patients with extremely low hemoglobin values. Specifically, the 5.1 to 6.0 g/dL range of Hb seemed to be the “sweet spot” where the real increase in mortality due to anemia manifested14. Only looking at mortality based on hemoglobin values, the 2.1-3.0 g/dL range resulted in a 50% mortality (n=6, 83.3% of whom underwent emergency procedures), the 3.1-4.0 g/dL 18.8% mortality (n=16, 43.8% of whom underwent emergency procedures), the 4.1-5.0 g/dL range 19.4% (n=31, 32.3% of whom underwent emergency procedures), the 5.1-6.0 g/dL range 14.3% mortality (n=49, 12.2% of whom underwent emergency procedures), and the 6.1-7.0 g/dL range with 5.2% mortality (n=58, 6.9% of whom underwent emergency procedures)14. Their adjusted OR for urgency, ASA score, and age was 1.82 for each 1g/dL decrease in Hb (95% CI 1.27-2.59, P=0.001)14.

In the Transfusion Requirements in Septic Shock (TRISS) trial, Holst et al15 studied ICU patients in septic shock with baseline hemoglobin values of 9 g/dL and randomized them into low threshold for transfusion groups (Hb of 7g/dL; n=502) and high threshold for transfusion groups (Hb of 9 g/dL; n=496) and then looked at 90 day mortality as the primary outcome measure. At 90 days 43% and 45% had died in the lower and higher-threshold groups, respectively (RR 0f 0.94, 95% CI 0.78-1.09; p=0.44), basically, non-statistically significant). Interestingly, however, the lower threshold patients received 50% less units of PRBCs15. Just think about the cost savings of not transfusing patients who essentially gain no benefit!

There are plenty more studies out there demonstrating a benefit to holding off on transfusion by utilizing a more restrictive approach. I chose to read up on the studies mentioned above because they looked at the so-called Hemoglobin-triggers for transfusion in pretty sick patients, either because of age, comorbidities, ICU or post-op settings, and not just any post-op, but hip fracture and cardiac surgery, so you know these are not your yearly marathon runners! But even in these patient populations transfusing at lower Hb thresholds was well tolerated and not that different in terms of mortality and morbidity compared to higher thresholds; in some cases, patients do even better, they are transfused less, and there are obviously cost savings in terms of resource utilization and hospital length of stay. It ultimately boils down to the question of what you are willing to accept: the danger of anemia or the danger of blood transfusion (see reference 5 for a great review on the perils of anemia versus transfusion)5.

A clinical guideline9 on the transfusion of blood can be helpful in this challenging decision. The AABB (American Association of Blood Banks) makes a strong recommendation based on high-quality evidence (n=6264 from 19 trials) that a restrictive 7 g/dL threshold for transfusion should be utilized for stable, hospitalized anemic patients, while 8 g/dL (based on the FOCUS trial) should be the trigger for post-op patients or when they manifest symptoms of anemia9. They make a weak recommendation of moderate quality evidence for transfusing stable, hospitalized patients with cardiovascular disease at a threshold of 8 g/dL or for symptoms of anemia, and they do not make any recommendations for or against transfusing acute coronary syndrome patients citing lack of clinical data from high-quality studies9.

After reading about this complex topic, I am left with more questions than answers, and a lot more room for clinical judgment in the decision to transfuse the euvolemic anemic patient that presents to the ED, even with clinical symptoms and signs of anemia. The literature seems to support a more restrictive “less is more” approach, perhaps pushing these patients a little bit more before pulling the transfusion trigger. What are your clinical approaches? How do you manage the patient with the non-acute or even acute anemia? And how do you decide when to transfuse? Do you think there should there be strict institutional protocols and clinical pathways that apply to broadly defined populations of anemic patients or should this be approached on a case-by-case basis?


Recommended Further Readings (not all papers, references 16-18, were cited in the blog post but were great reads nonetheless):

1)Villanueva C, Colomba A, Bosch A, et al. Transfusion strategies for acute gastrointestinal bleeding. N Engl J Med. 2013;368(1):11-21.

2)Goodnough LT, Levy JH and Murphy MF. Concepts of blood transfusion in adults. The Lancet. 381:1845-1854.

3)Goodnough LT. Blood management: transfusion medicine comes of age. The Lancet. 2013;381: 1791-1792.

4)Marino PL. The ICU Book. Philadelphia: Wolters Kluwer Health/Lippincott Williams & Wilkins. 2014, Print.

5)Shander A, Javidroozi M, et al. What is really dangerous: anemia or transfusion? British Journal of Anesthesia. 2011; 107(S1): i41-i59.

6)Scott K, Greineder C, and Weinberger Conion L. The Use of Blood Products In The Critically Ill Patient: Indications and Risks. EM Critical Care in EB Medicine. 2014;4(1): 1-20.

7)Lelubre C and Vincent J. Red blood cell transfusion in the critically ill patient. Annals of Intensive Care. 1(43): 1-9.

8)US Food and Drug Administration Center for Biologics Evaluation and Research. Fatalities reported to FDA following blood collection and transfusion. Annual summary for fiscal year 2013. (accessed Aug 8, 2015).

9)Carson JL, Grossman BJ, et al. Red Blood Cell Transfusion: A Clinical Practice Guideline From the AABB. Annals of Internal Medicine. 2012; 157:49-58.

10)Hebert PC, et al. A multicenter, randomized, controlled clinical trial of transfusion requirements in critical care, N Engl J Med. 1999;340(6):409-417. (TRICC)

11)Carson JL, Sieber F et al. Liberal versus restrictive blood transfusion strategy: 3-year survival and cause of death results from the FOCUS randomized controlled trial. The Lancet. 2015;385:1183-1189.

12)Carson JL, Terrin ML, et al. Liberal or Restrictive Transfusion in High-Risk Patients after Hip Surgery. N Engl J Med. 2011; 365(26): 2453-2462. (FOCUS)

13)Hajjar LA, Vincent J et al. Transfusion Requirements After Cardiac Surgery, The TRACS Randomized Controlled Trial. 2010; 304(14):1559-1567. (TRACS)

14)Shander A, Javidroozi M, et al. An update on mortality and morbidity in patients with very low postoperative hemoglobin levels who decline blood transfusion. 2014;54:2688-2695.

15)Holst LB, Haase N et al. Lower versus Higher Hemoglobin Threshold for Transfusion in Septic Shock. N Engl J Med. 2014;371(15): 1381-1391. (TRISS)

16)Ansari S and Szallasi A. Blood management by transfusion triggers: when less is more. Blood Transfus 2012;10:28-33.

17)Carson JI and Hebert PC. Should We Universally Adopt a Restrictive Approach to Blood Transfusion? It’s All About the Number. The American Journal of Medicine. 2014; 127(2): 103-104.

18)Janz TG, Johnson RL, and Rubenstein SD. Anemia In The Emergency Department: Evaluation And Treatment. Emergency Medicine Practice in EB Medicine. 2013; 15(11): 1-16.

The following two tabs change content below.


Latest posts by ray (see all)

Leave a Reply

Your email address will not be published.