Bruce D. Spiess, MD, FAHA
Introduction
Blood transfusion has evolved into a mainstay medical therapy. It has never undergone efficacy and safety testing the way in which a drug would be tested for approval by the US Food and Drug Administration.[1,2] Today, some 107 years after Karl Landsteiner discovered the ABO histocompatibility antigens, we are still learning about risks and outcomes of transfusion. There have been relatively few randomized trials of transfusion,[3,4] but the lay public and many physicians are aware of several concepts. Fourteen to 15 million units of red blood cells are transfused per year in the United States to 4.5 million patients. Some patients die of complications from their transfusions. It is true that today we have the "safest" blood; however, this term is misunderstood.
Infection Risks
The 2 major viral concerns (hepatitis C and HIV/AIDS) of the 1940s-80s are no longer major risks of transfusion.[5,6] Through multiple layers of safety monitoring, the blood-banking community has made hepatitis C, HIV/AIDS, and West Nile virus a vanishingly small risk (1 in 2.4 million units or greater).[5,6] Hepatitis B, parvovirus, Epstein-Barr virus, cytomegalovirus (CMV), emerging viruses (Avian flu and SARS), as well as others are still of concern, of which hepatitis B, Epstein-Barr virus, and CMV are relatively common.[6]
The emerging viruses pose an immeasurable and as yet small risk. There are possibilities of contracting parasitic pathogens not presently eliminated from our blood supply (Chagas disease, etc.).[6,7] Malaria remains the number-one communicable disease risk of transfusion for mankind, but in the industrialized world that risk is small indeed.[6,8] Mad cow disease (variant Creutzfeldt-Jakob disease or vCJD in humans) can be transmitted through transfusion; presently no tests are available to detect the abnormal proteins (prions) that cause the disease, but the risk is far smaller than even that of HIV/AIDS.[6,9]
Bacterial contamination of units of blood is a possible risk, but the blood-banking community is ever vigilant.[6,10] Since the risks of HIV/AIDS and hepatitis C have been so dramatically reduced it is correct to think of our blood supply as the "safest." However, blood transfusion is inherently unsafe; it is accepted that some disease and deaths will occur from the act of transfusion.
Effect on Recipient's Immune System
The transfusion of banked blood from a donor to a recipient is a liquid organ transplant. Through that transplant, the recipient's own immune system is altered for some period of time.[11] In the medical world there is debate in regard to how long and how important that immune suppression/alteration lasts; likely it lasts for weeks and perhaps up to years. In the last year we have learned that chimerism (multiple cell lines in the recipient trying to identify foreign cells from other sources) may well be responsible for some aspects of autoimmune diseases such as arthritis, lupus, and others.[12] What are the long-term effects of blending DNA in an individual and in a population? Blood bankers are constantly debating this issue. Only years of future research will tell us how important the effect of chimerism is on our human health.
Human Error
Human error in medicine is a recognized fact.[13,14] Sixty or more human interactions must occur for a patient to receive a unit of blood. The human events begin from the moment that blood is collected from a donor or that a patient has his/her blood drawn for blood typing until it is infused into a recipient. Human error in this scenario can include mislabeling, using the wrong unit, or confusing one patient's blood sample with another patient's sample.
Each human touch is a source of potential disastrous consequences. Between 1 in 6000 and 1 in 20,000 units of blood transfused result in a patient receiving the wrong unit of blood.[15,16] Most of the time the effects are minor, but a patient dies from massive hemolytic transfusion crisis every 1 in 100,000 units transfused. Far more often (perhaps 10 times more often) there is a minor delayed hemolytic event indicating that a minor antibody/antigen reaction (untested) has occurred. We only test for a very small minority of blood antigens in a cross-match. The red cells transfused are broken apart and free hemoglobin is released, which in itself has profound consequences. Medical care today does not regularly monitor for this phenomenon.
Transfusion-Related Acute Lung Injury
Transfusion-related acute lung injury (TRALI) is a risk of transfusion [17,18] and possibly the most common and most serious risk. TRALI, for which the international definition of was agreed upon 4 years ago, is defined as new acute lung injury (ALI) that occurs during or within 6 hours of transfusion, not explained by another ALl risk factor.[19] It is caused by the donor's antibodies attacking the recipient's endothelial cells lining the capillaries of the lung vasculature. There are considerable arguments about how often this event happens. TRALI is a diagnosis of exclusion, meaning that doctors first examine other reasons why the lungs could be abnormally filled with fluid rather than assigning transfusion as the cause.
Today we accept that the risk for TRALI is 1 in 2000-4000 units transfused.[19] However, most patients who receive a transfusion receive multiple units of blood. One study that investigated incidence of TRALI and transfusion-associated circulatory overload (TACO) in critically ill patients found that the incidence of possible TRALI was 1 in 534 units transfused; suspected TRALI, 1 in 1271 per unit transfused; and TACO, 1 in 356 per unit transfused.[17] This is a very common serious adverse event, especially when one compares these TRALI data with that for HIV/AIDS and hepatitis C (1 in 2.4 million units). In studies in which the highest risk for TRALI was found, it was noted that TRALI was associated with 50% mortality.
Transfusion and Patient Outcomes
A large number of recent (within the last 7-10 years) studies have shown a profound association between transfusion and adverse outcomes. Very few articles have been published showing a relationship between transfusion and improved outcomes. Today there is mounting evidence that patients who are transfused do worse than patients who are not.[20]
In orthopaedic surgery, patients who have received allogeneic (banked) blood have a 1.5- to 3.5-fold increased risk for perioperative major infection.[21-23] This includes wound infection, pneumonia, sepsis, and urinary tract infection. In hip or knee replacement surgery, if an infection develops, there are serious consequences: prolonged hospital stay, removal of the artificial joint, and the possibility of never walking normally again.
In open heart surgery, the same data are available from many sources.[24-28] Those who receive units of blood have a longer stay in the hospital, have more chance for pneumonia, more infections (including major wound infections), more respiratory complications, more kidney failure, more heart attacks, more strokes, more returns to the operating room for bleeding, and ultimately at least twice the death rate than those not transfused. Patients who are transfused have more heart arrhythmias (atrial fibrillation), heart failure, inability to wean from the bypass machine, and other additional adverse outcomes.[29]
Over the long term (10 months to 5 years), the death rate of patients transfused in the hospital for heart surgery is profoundly worse (at least double) compared with those not transfused. Are these associations a result of cause and effect, or just a reflection of the fact that patients who are more ill and are undergoing more complex surgery receive more blood transfusions? Statistical analysis performed on the data to control for confounding variables — such as patient risk or length of surgery, — have been equalized through multivariate and propensity statistics, yet the effect of transfusion stands out as an independent and highly significant association. Even with the finest statistical control of potential confounders, cause and effect cannot be claimed. It fits the biology of transfusion, but we do not know conclusively.
Conclusion
So, how will we ever know about cause and effect? The true risks and benefits of allogeneic (donor blood) transfusion will be debated by academic physicians. Today, with HIV/AIDS and hepatitis C no longer a major risk of transfusion, experts can focus on other risks that were always present but were underinvestigated. The world of blood transfusion is constantly changing. The societal perception is that "transfusion is good'; we do not know that for many patients.
