Blood carries out the vital functions of human life. It is composed of cells, suspended in plasma, that carry oxygen and nutrients to organs, remove carbon dioxide and waste products for excretion, fight infection and aid wound-healing. When blood loss is severe - such as due to trauma or surgery - it is considered necessary to replace blood components to avoid certain death. This is usually been done via blood transfusion, where whole blood or its components (plasma, blood cells or platelets) is administered into the veins via a drip.
Is it safe?
No. The dangers of transfusing blood are well known, and it has been associated with a high mortality rate ever since it was first introduced in the 17th century. Today, although the associated death rate is significantly lower, the procedure still carries considerable risks.
Blood types are based on the antigens (proteins) present on the surface of red blood cells. The three antigens (A, B and O) give rise to four blood groups: A, B, AB and O. Another important protein is the Rhesus (Rh) factor, so the four groups are further classed as Rhesus positive (Rh1) or negative (Rh2). Transfusing one type of blood into a person of a different blood type can lead to serious reactions, including death.
Moreover, there is a relatively high incidence of non-fatal adverse reactions. Around 5 per cent of transfused patients have a reaction, with symptoms that may be hard to distinguish from their original condition. Milder symptoms include chills and fever, aches and pain, including back pain, and hives and itching.
Major reactions include immunological complications. Red cell alloantibodies can cause red cell destruction (BMJ, 2002; 325: 143-7), leading to shortness of breath, severe headaches, chest or back pain, blood in the urine and even death. In 2003, 25 cases of such delayed haemolytic reactions to transfusion were reported, according to SHOT (Serious Hazards of Blood Transfusion), a UK medical inquiry group affiliated with the NHS. These reactions are believed to be underrecognised as, often, there are no symptoms, or the clinical signs are subtle.
Also, the recipient's blood may clot, or the recipient may have an allergic reaction that could lead to anaphylactic shock. Other potentially dangerous complications include:
* a viral disease. Worldwide, it is thought that 5-10 per cent of cases of HIV are transmitted through blood transfusions. In the UK in 2003, there were eight cases of transfusion-transmitted viral infections, including two cases of hepatitis B, and one case each of HIV, hepatitis A and malaria (SHOT Annual Report, 2003).
* a bacterial infection. Blood must be stored at room temperature (20-24 degrees C) and so creates a perfect environment for bacterial growth. Studies show that one in 1000-2000 platelet units are bacterially contaminated. In developed countries, contaminated platelets are the most common cause of death among transfusion-related complications, with an estimated risk of death as high as one in 100,000 (Curr Hematol Rep, 2004; 3: 121-7).
* long-term mild immunosuppression, which can result in poorer recovery, and a further risk of infection or malignancy (BMJ, 2002; 325: 143-7).
* acute lung injury, post-transfusion purpura (bleeding into the skin causing a rash) and transfusion-related graft-versus-host disease (BMJ, 2002; 325: 143-7). Of 36 cases of suspected transfusion-related acute lung injury, nine patients died, with one case definitely, and seven possibly, due to transfusion. Another 22 suffered short-term major illness and five minor illnesses. In 21 cases, there was confirmed leucocyte incompatibility due to plasma-rich blood components in all but one patient (SHOT Annual Report, 2003).
* congestive heart failure from a transfusion that is too rapid.
* acute kidney failure.
What doctors don't tell you
Human error in the transfusion process is an important cause of adverse reactions. The most common mistake - constituting 70 per cent of the human errors reported from 1999 to 2000 - is giving the wrong blood product. Giving an incompatible blood type was reported 97 times during 1999-2000, and caused four deaths and 29 cases of "immediate major morbidity". Fatal misidentification errors are thought to occur once every 600,000- 800,000 transfusions, and non-fatal errors once in 12,000-19,000 transfusions (BMJ, 2002; 325: 143-7).
Worryingly, these figures are increasing, with multiple errors now contributing to just over half of all 'wrong blood' events. The most common error is not carrying out a pretransfusion bedside check (SHOT Annual Report, 2003).
Blood contains many more antigens than previously thought that can cause adverse reactions. In theory, if a patient is given the correctly matched blood, the procedure should be safe. The problem is, there are some 400 other antigens in the blood that may also trigger a reaction.
Not all pathogens are screened for. In theory, blood transfusions are safe provided that the blood types are correctly matched and the blood is free of disease-causing organisms. However, practical screening techniques are not available for all pathogens, such as those causing malaria, Chagas disease (American trypanosomiasis, caused by the parasite Trypanosoma cruzi), West Nile disease and variant Creutzfeldt-Jakob disease (vCJD), or such screening is considered to be not cost-effective (Curr Opin Hematol, 2003; 10: 412-8).
Animal (sheep) studies provide some evidence that bovine spongiform encephalopathy (BSE or mad-cow disease) can be transmitted by blood transfusion (BMJ, 2002; 325: 143-7). The first possible case of vCJD transmission was reported in 2003, following the death of a transfusion recipient (SHOT Annual Report, 2003).
Where it is not possible to screen for pathogens, the 'honour system' is relied upon. This means that potential donors must state whether or not they have been exposed to, or are a carrier of, an infectious disease. This includes instances where a donor has travelled to a region where an infectious disease like malaria or babesiosis (a life-threatening parasitic infection) is prevalent. Such a system is intrinsically unreliable because the donor may either be unwilling or unable to admit they are at risk.
Blood transfusions may have a negative impact on the survival of cancer patients. In a study of surgical patients with colorectal cancer, the risk of recurrence in transfused patients - whether with matching (allogeneic) or their own (autologous) blood - was about twice as high as in non-transfused patients (N Engl J Med, 1993; 328: 1372-6). Similarly, a study of cancer of the head and neck found that transfused patients were twice as likely to have recurrences, or die, as non-transfused patients (Laryngoscope, 1994; 104: 95-8). In cases of cancer of the mouth, throat, nose or sinuses, the recurrence rate was 31 per cent without, and 71 per cent with, a blood transfusion (Am J Surg, September 1986).
Tests for blood-borne diseases are not infallible. There is still a risk of contracting a disease that has been screened for although, for most diseases, the chances are thankfully very small. The risk of contracting HIV through a blood transfusion in the UK is one in eight million. The statistics for hepatitis B are higher, running at around one in 260,000 (Vox Sang, 2003; 84: 274-86).
Some bloodborne diseases do not develop detectable antibodies and so can be missed by screening. Chagas disease and vCJD may not be detectable for several years after infection. HIV antibodies don't develop for several months.
Blood transfusions can encourage more bleeding by preventing bloodclotting. In 1986, the British Journal of Surgery reported that, before transfusions became commonplace, the gastrointestinal haemorrhage mortality rate was 2.5 per cent. Now, several studies have reported a death rate of 10 per cent, possibly because transfusion reverses the blood-coagulation process - the body's natural response to haemorrhage - and encourages rebleeding.
Blood transfusions can transmit diseases that are otherwise not considered infectious, such as leukaemia and lymphomas (Postgrad Med J, 2001; 77: 230-4).
In the US, commercial blood banks are five to eight times more likely to be located in areas that are considered 'high risk' in terms of the incidence of behaviour associated with transfusion-transmissible infections.
A study that monitored the distribution of blood banks in the period between 1980 and 1995 found that they are overrepresented in areas with "very active local drug economies" - a pattern that has remained even after the links between HIV and hepatitis C and plasma products were clearly established, and which has obvious implications for blood safety in the US (Am J Public Health, 2004; 94: 1224-9).
Red blood cell (RBC) transfusions don't necessarily improve the recovery chances of the critically ill. A recent study by the University of Maryland found that such transfusions do not consistently improve oxygen consumption either throughout the body tissues or at the level of cellular circulation; it may also result in a worse outcome for those who are severely ill. Indeed, it is difficult to identify those who would improve with RBC transfusion (Crit Care Clin, 2004; 20: 255-68).
Children are especially at risk. Of the 348 cases of incorrect blood-component transfusion reported in the UK (SHOT Annual Report, 2003), 53 involved patients under 18 years of age, with 28 of those being less than one year old and 71 per cent (20 cases) being infants in their first month of life. Nine of these paediatric patients became potentially or actually ill - in five cases due to haemolysis (red cell destruction) and, in two cases, due to RhD2 babies being given RhD1 blood.
Blood transfusions that are correctly cross-matched may still be rejected. A blood transfusion is effectively a tissue transplant, and a recipient's immune system may sense the transfused blood and reject it in exactly the same way as it would a transplanted heart, liver or other organ.