Appearance
13.7 — Blood Groups and Transfusion
Before 1900, transfusion was a lottery. Some patients recovered dramatically; others died within minutes, and nobody could predict which. Karl Landsteiner solved it in 1901 by mixing blood samples from his colleagues and observing which combinations clumped. He found three groups; a fourth was identified shortly afterwards. He received the Nobel Prize in 1930.
It is a rare case of a Nobel-winning discovery that could have been made a century earlier with equipment available at the time. What was missing was the idea that people's blood might differ systematically.
The ABO system
The antigens are sugars, not proteins, attached to lipids and proteins on the red cell surface — part of the glycocalyx of Chapter 1.4.
Everyone starts with the same base sugar structure, called H antigen. An enzyme then adds one further sugar to it.
The A gene codes for an enzyme that adds one sugar. The B gene codes for a slightly different enzyme adding a different one. The O gene codes for a non-functional enzyme, so nothing is added.
So O is not a different antigen — it is the absence of an addition. The two enzymes differ by only a handful of amino acids, and the O allele is typically a single-base deletion producing a non-functional protein (Chapter 2.5).
A and B are codominant; O is recessive (Chapter 2.6).
| Group | Antigens on cells | Antibodies in plasma | Frequency (varies by population) |
|---|---|---|---|
| A | A | anti-B | ~40% |
| B | B | anti-A | ~10% |
| AB | A and B | none | ~4% |
| O | none | anti-A and anti-B | ~45% |
The pattern is the key: you have antibodies against whichever antigens you lack.
And that is genuinely unusual. For every other antigen, you only make antibodies after exposure. ABO antibodies are present from a few months of age without any transfusion.
They arise because gut bacteria carry sugar structures closely resembling A and B antigens. The immune system responds to them, and the resulting antibodies happen to cross-react with red cells. So the exposure is real; it simply comes from the gut rather than from blood.
They are IgM (Chapter 13.3) — pentameric, ten binding sites, and outstanding at agglutination and complement activation.
Which is exactly why an ABO-incompatible transfusion is so catastrophic and so fast.
The universal donor and recipient
O negative is the universal donor — no A, B or D antigens, so nothing for the recipient's antibodies to attack.
AB positive is the universal recipient — no anti-A or anti-B antibodies, so any red cells are accepted.
And for plasma it is exactly reversed. Plasma contains antibodies rather than antigens, so AB plasma is the universal donor — it contains neither anti-A nor anti-B — and O plasma can only be given to O recipients.
This reversal catches people out constantly, and it follows directly from asking which component carries the antibodies.
The Rh system
More than 50 antigens, of which D is by far the most important — so "Rh positive" means D positive.
Around 85 percent of people of European descent are Rh positive, with substantial variation between populations.
And unlike ABO, anti-D antibodies are not present naturally. They only appear after exposure to D-positive red cells, which happens through transfusion or pregnancy.
Anti-D is IgG, and that has one crucial consequence: IgG crosses the placenta (Chapter 13.3).
Haemolytic disease of the newborn
This is the classic and most instructive example of maternal–fetal immune conflict, and it has been almost entirely eliminated.
The setup: an Rh-negative mother carrying an Rh-positive baby — which requires the father to be Rh positive.
First pregnancy: the two circulations are separate, so exposure is minimal until delivery. At delivery, fetal red cells enter the maternal circulation, and the mother makes anti-D. The first baby is unaffected, because sensitisation happens at the end.
Subsequent pregnancies: if the baby is again Rh positive, maternal anti-D crosses the placenta and destroys fetal red cells.
The consequences range from mild anaemia and jaundice to severe anaemia with heart failure and fluid accumulation throughout the fetus — hydrops fetalis — which was frequently fatal.
And there is a specific danger after birth. Massive red cell breakdown produces large amounts of bilirubin (Chapter 7.1). The placenta was clearing it; after delivery the immature newborn liver cannot keep up. Very high bilirubin crosses into the brain and causes kernicterus — permanent damage to the basal ganglia, producing cerebral palsy and deafness.
The prevention is elegant and almost completely effective.
Give the Rh-negative mother anti-D immunoglobulin — antibodies against D — at around 28 weeks and within 72 hours of delivery, and after any event that could cause fetal–maternal bleeding such as miscarriage, amniocentesis or abdominal trauma.
The given antibody binds and clears any fetal red cells before the mother's own immune system can respond to them.
So you prevent an immune response by supplying the antibody yourself. It is deliberately using passive immunity to prevent active immunity (Chapter 13.3), which is a genuinely counter-intuitive use of the concept.
The result: haemolytic disease of the newborn has fallen by over 95 percent where the programme is implemented. It went from a leading cause of stillbirth and neonatal brain damage to a rarity within a generation.
And fetal Rh type can now be determined from fetal DNA circulating in maternal blood (Chapter 2.9), so anti-D can be given only to the mothers who actually need it — avoiding unnecessary blood product exposure in around 40 percent.
ABO incompatibility can also cause haemolytic disease — most often an O mother with an A or B baby — but it is much milder, because ABO antibodies are mainly IgM and cross the placenta poorly, and because A and B antigens are less well developed on fetal cells.
Other blood group systems
More than 40 systems and over 600 antigens are recognised.
Kell, Duffy, Kidd and MNS are the ones that matter clinically, because antibodies against them can cause transfusion reactions and haemolytic disease.
And Duffy has a striking evolutionary story. The Duffy antigen is the entry receptor used by Plasmodium vivax, one of the malaria parasites. Around 95 percent of people of West African descent are Duffy negative — they lack the receptor entirely — and are therefore essentially resistant to vivax malaria.
It is another malaria-driven selection story alongside sickle cell and G6PD deficiency (Chapter 3.3), and it is why vivax malaria is nearly absent from West Africa.
Antibodies against these minor antigens develop after exposure, which is why people who have had multiple transfusions — those with sickle cell disease or thalassaemia — become progressively harder to match. They are typed extensively and matched for more than ABO and D from the start, precisely to avoid this.
Cross-matching
Three steps before any transfusion.
ABO and Rh typing — testing the patient's cells against known antibodies, and their plasma against known cells. The two must agree, and disagreement is a stop signal.
Antibody screen — testing the patient's plasma against a panel of cells carrying known antigens, to find any antibodies from previous transfusion or pregnancy.
Cross-match — mixing the actual donor unit with the patient's plasma. The final check.
In an emergency, O negative blood can be given immediately; a group-specific unit takes about 10 minutes; a full cross-match takes 30 to 45 minutes. The choice depends on how fast the patient is bleeding.
And the single most important safety measure is not laboratory work at all.
The commonest cause of a fatal transfusion reaction is giving the right blood to the wrong patient. A clerical or identification error, not an immunological one.
Which is why the bedside check is done by two people, at the bedside, immediately before starting, with the patient's identity confirmed verbally where possible. It feels like bureaucracy and it is the step that prevents deaths.
Transfusion reactions
Acute haemolytic reaction — the ABO-incompatible catastrophe.
Complement is activated massively, red cells lyse inside the vessels, and the consequences cascade: fever and rigors, chest and back pain, a sensation of impending doom, low blood pressure, dark urine from free haemoglobin, kidney failure, and disseminated intravascular coagulation (Chapter 7.1).
It can be fatal within minutes, and it can begin after as little as 10 to 15 ml.
Which is why the first 15 minutes of any transfusion are observed closely — the reaction, if it is coming, usually declares itself then.
Treatment: stop the transfusion immediately, keep the intravenous line open with saline, and support the circulation and kidneys.
And a specific point about conscious versus unconscious patients. An awake patient reports pain and a feeling that something is badly wrong, which is often the earliest sign. An anaesthetised patient cannot — so unexplained low blood pressure or bleeding during surgery in a patient receiving blood is treated as a reaction until disproved.
Febrile non-haemolytic reaction — fever and chills without haemolysis, from antibodies against donor white cells. Common, unpleasant, not dangerous. Largely prevented by leucodepletion — removing white cells from all donated blood — which is now routine in most countries.
Allergic reactions — from plasma proteins. Usually mild urticaria; occasionally anaphylaxis.
And there is one specific and severe version worth knowing. People with IgA deficiency can have severe anaphylaxis to transfused plasma containing IgA, because they may have anti-IgA antibodies. They need washed or IgA-deficient products.
Transfusion-related acute lung injury (TRALI) — sudden severe breathlessness and lung oedema within six hours, from donor antibodies reacting with recipient white cells in the lung.
It was a leading cause of transfusion-related death, and the incidence has fallen substantially since blood services began preferentially using plasma from male donors — because the antibodies responsible are largely generated during pregnancy. A change in donor selection, based on understanding the mechanism.
Transfusion-associated circulatory overload (TACO) — simply too much volume too fast, causing heart failure. Now recognised as one of the commonest serious complications, particularly in the elderly and in those with heart failure, and largely preventable by transfusing more slowly and one unit at a time.
Delayed haemolytic reaction — days to weeks later, from an antibody against a minor antigen that was too low to detect at cross-match and has been boosted by the transfusion. Usually mild.
Infection — and this is where the picture has improved most dramatically.
Blood safety
The contaminated blood disasters of the 1970s and 1980s are essential context and should not be skipped.
Thousands of people with haemophilia were infected with HIV and hepatitis C through contaminated clotting factor concentrates, which were pooled from thousands of donors, so a single infected donation contaminated an entire batch (Chapter 2.8). Many died. Public inquiries in several countries have found systemic failures.
What followed was a transformation in blood safety, and the numbers now are remarkable.
Donor selection and deferral. Testing every donation for HIV, hepatitis B, hepatitis C, syphilis and others. Nucleic acid testing detects viral genetic material rather than waiting for antibody, which shortens the window period after infection during which a donor tests negative — from months to days for HIV.
Current estimated risk per unit in high-income countries: HIV around 1 in 2 million or lower; hepatitis C around 1 in 2 million; hepatitis B around 1 in 1 million.
Pathogen reduction technologies treat platelets and plasma to inactivate any organisms present.
And the residual concerns are the ones that testing cannot address: emerging infections not yet tested for, and variant CJD, the prion disease (Chapter 1.1), which cannot be tested for at all — hence permanent deferral of donors who may have been exposed, and leucodepletion.
Using blood well
And the most important development in transfusion medicine over the last two decades is that less is better.
Restrictive transfusion thresholds have been tested repeatedly and outperform liberal ones. In most stable patients, transfusing at a haemoglobin of 70 g/L produces outcomes as good as or better than transfusing at 90 to 100.
This surprised almost everyone, and it has changed practice worldwide. Transfusion carries risks, and giving blood that is not needed adds risk without benefit.
"Patient blood management" is the resulting approach, and it has three arms:
Treat anaemia before elective surgery — iron, B12, folate, and erythropoietin where relevant. A patient who arrives for surgery with a normal haemoglobin needs far less blood.
Minimise blood loss — surgical technique, tranexamic acid, and cell salvage, in which blood lost during surgery is collected, washed and returned.
Optimise tolerance of anaemia rather than reflexively correcting the number.
Tranexamic acid deserves specific mention. It blocks clot breakdown, and in the CRASH-2 trial it reduced death from bleeding in trauma patients — a cheap, off-patent, widely available drug with a measurable mortality benefit. The effect is greatest when given within an hour and disappears after three hours, which makes early administration a genuine priority.
The components
Whole blood is rarely used. One donation is separated into components, and each goes to a patient who needs that specific part.
Red cells — for anaemia and blood loss. Stored 35 to 42 days at 4 °C.
Platelets — for low counts or platelet dysfunction. Stored at room temperature with agitation for only 5 to 7 days, which makes them the hardest component to keep in stock, and the room temperature storage is why they carry the highest bacterial contamination risk.
Fresh frozen plasma — clotting factors. Frozen for up to a year.
Cryoprecipitate — concentrated fibrinogen and specific factors.
One donation therefore helps up to three patients, which is the practical argument for component separation.
What the next page fixes
If the immune system rejects red cells over a single sugar, an entire organ from another person is a far larger problem. Chapter 13.8 covers transplantation — what is actually being rejected, how it is prevented, and why a transplanted kidney is nonetheless one of the most successful operations in medicine.