In 1900, a young Austrian pathologist took blood samples from the staff of his own institute. When he mixed some of those samples together, they clumped.
Some of the colleagues he had sampled were already dead. They had died on the operating table from blood transfusions that should have been compatible.
 |
| Landsteiner with the Nobel medal he was awarded in 1930 for the ABO work. The original version of this post did not mention the prize. Bachrach Studios, public domain, via Wikimedia Commons |
 |
| Landsteiner at work. He was still a fairly junior researcher when he did the ABO work, and the finding was rejected by several journals before it was accepted. Robert Prummel, CC BY-SA 3.0, via Wikimedia Commons |
Karl Landsteiner worked out why. He was awarded the Nobel Prize in Physiology or Medicine in 1930 for it, and the system he named is still the first thing written on a transfusion request.
The earlier version of this post ran to about 870 words. It never mentioned the deaths, it never mentioned the Nobel, and it contained two claims that could cause real harm if acted on.
How the system works
A blood group is decided by two things sitting in different places. On the outside of the red blood cell are antigen molecules. In the plasma are antibodies directed against the antigens that are not there.
An antigen is any substance the immune system can recognize and respond to. An antibody is a protein made to fit a particular antigen closely enough to grab it.
 |
| Red blood cells under an electron microscope. Each is a flexible disc with no nucleus, which is what lets it squeeze through a capillary narrower than itself. The antigens that define a blood group are molecules on the outside of that membrane. Bruce Wetzel and Harry Schaefer, public domain, via Wikimedia Commons |
The rule is worth stating exactly, because the whole system falls out of it: a person makes antibodies against the antigens they do not have. Never against the ones they do.
That is not a strange choice. It is what keeps a person from attacking their own cells. The immune system learns what "self" looks like, and anything matching it is left alone.
Why clumps are dangerous
When an antibody meets its antigen on a red cell, it bridges two cells together. More antibodies bridge more cells, and the result is agglutination: visible clumping.
 |
| An antibody molecule. The Y shape is the point: two identical arms that can bind an antigen, joined by a stem that tells the rest of the immune system what to do. Anti-A and anti-B differ in the shape of the arm tips. Tokenzero, CC BY-SA 4.0, via Wikimedia Commons |
Clumps are a mechanical problem as much as an immune one. They are too large to pass through a capillary, so they block small vessels, and the cells inside them break down rather than doing their job.
An incompatible transfusion is therefore dangerous in two ways at once: it damages the tissue the cells were meant to reach, and the free haemoglobin released is itself toxic.
The four ABO groups
| Blood group | Antigen on red cell | Antibody in plasma |
| A | A | Anti-B |
| B | B | Anti-A |
| AB | A and B | None |
| O | None | Anti-A and Anti-B |
The ABO system. Every group follows from the rule that you make antibodies against the antigens you lack.
Group O is the extreme case in both directions. Its red cells carry nothing to attack, and its plasma carries both antibodies. Group AB is the opposite extreme: its red cells carry both antigens, and its plasma carries neither antibody.
How blood group is inherited
The ABO gene sits on chromosome 9. It comes in three common versions, conventionally written I superscript A, I superscript B, and plain i.
The two capital versions are codominant, which means a person carrying both makes both antigens and is therefore group AB. The lowercase version is recessive to both.
| Genotype | Blood group |
| IAIA or IAi | A |
| IBIB or IBi | B |
| IAIB | AB |
| ii | O |
Six genotypes, four blood groups. The distinction matters for predicting a child.
A person with group A blood could be IAIA or IAi. Those look identical on a blood test and behave identically in a transfusion, but they predict completely different children.
Working a cross
Take a group A father and a group B mother. Assuming the commonest genotypes, IAi and IBi, there are four equally likely children.
| i | i | |
| IA | IAi | IAi |
| i | ii | ii |
A cross of IAi by IBi. Two cells are group A, one is group AB, and one is group O.
So the possible children are A, A, AB and O. Each is equally likely before you know anything else about them.
Run the case the other way and the logic is sharper. A group B mother and a group O child: the child must be ii, so the mother must have passed on an i, which means she is IBi.
A group A father of type IAIA cannot have a group O child, because he has no i to give. A father who is IAi could.
A blood group does not change through life, with one rare exception: a bone marrow transplant can change a person's blood group, because the donor's cells carry the donor's antigens.
Transfusion, and the donor rule the original got wrong
This is where the earlier version of this post was most misleading, so it is worth being precise.
The universal donor for red cells is O negative. Not O. The word negative carries real clinical weight.
An O negative red cell has no A antigen, no B antigen, and no D antigen, so there is nothing on it for a recipient's plasma to attack. An O positive red cell does carry the D antigen, and giving one to an Rh-negative woman of childbearing potential can sensitise her against Rh.
Plasma runs the opposite way, and this is the part the original left out. Type O is the worst plasma donor there is, because O plasma contains both anti-A and anti-B.
The universal plasma donor is AB, because AB plasma contains neither antibody. The donor and recipient rules are mirror images, which is the elegant part: what makes a good red cell donor makes a bad plasma donor, and the reverse.
| What is being given | Universal donor | Why |
| Red cells | O negative | No A, B or D antigen to be attacked |
| Plasma | AB | No anti-A or anti-B to attack the recipient |
Donor and recipient rules run opposite ways. "Universal" always refers to one of these, never to a whole unit of blood.
There is a hard limit even on O negative. A person who is O negative can only be given O negative red cells, because their plasma contains anti-A and anti-B and would destroy anything else.
 |
| What agglutination looks like. Each antibody bridges two cells, and the clumps are too large to pass through a capillary. Mikael Haggstrom, public domain, via Wikimedia Commons |
 |
| A transfusion in progress. Blood is matched twice: once by group on the label, and again by an individual crossmatch against the patient, because blood group alone is not considered enough. BruceBlaus, CC BY 3.0, via Wikimedia Commons |
 |
| Symptoms of an acute hemolytic transfusion reaction. Fever, chills and back pain are the early signs, and this is why a transfusion is monitored closely: it is a medical emergency. Mikael Haggstrom, public domain, via Wikimedia Commons |
The Rh factor
The Rh system is a separate classification that sits on top of ABO. It turns on a single antigen, called D.
The name comes from a surprise. Landsteiner and Wiener, working around 1940, injected rhesus macaque blood into people and found that it agglutinated. They named the antigen after the animal.
 |
| A rhesus macaque. The Rh group is named for these animals because the D antigen was first found on their red cells. Nothing about a human Rh factor comes from the monkey; only the name does. Anvesha456, CC BY-SA 4.0, via Wikimedia Commons |
| Genotype | Rh type |
| DD or Dd | Rh positive |
| dd | Rh negative |
Two alleles. D is dominant, so one copy is enough to be Rh positive.
One important difference from ABO. Anti-A and anti-B occur naturally in everyone, without any exposure. Anti-D does not.
An Rh-negative person makes anti-D only after being exposed to Rh-positive blood. That is what makes sensitisation possible at all.
It is also why the original post was wrong to treat every Rh antibody this way. The statement holds for anti-D, but some other anti-Rh antibodies do occur naturally.
Rh in pregnancy
An Rh-negative mother carrying an Rh-positive fetus may, after exposure, start making anti-D. Those antibodies can cross the placenta and destroy fetal red cells.
The condition is called hemolytic disease of the newborn, and in its older name erythroblastosis fetalis. It is a problem for subsequent pregnancies far more than the first, because the mother has to be exposed before she can respond.
Destruction of fetal red cells causes severe anaemia, and the bilirubin released causes jaundice. The neurological injury from that bilirubin has a name, kernicterus, and it is the specific thing being prevented. Untreated, the condition can kill the fetus or the newborn.
Prevention is immunoglobulin. An Rh-negative pregnant woman is given anti-D immunoglobulin.
It mops up any fetal Rh-positive cells that have crossed into her circulation, before her immune system has a chance to respond to them.
It is given around the 28th week of pregnancy, and again within 72 hours of delivery, and also after any sensitising event such as a miscarriage, trauma, or an amniocentesis.
Worth noting what the original implied and should not have: ABO mismatch can also cause hemolytic disease of the newborn. It is generally milder than the Rh form, but it is not an Rh-only condition.
A worked case
An Rh-negative woman and an Rh-positive man whose own father was Rh-negative. The father must be Dd, because his father could only pass on d.
Crossing dd with Dd gives half Dd and half dd, so the chance of an Rh-positive child is 50 percent, and each pregnancy is independent.
Two Dd parents give a 25 percent chance of an Rh-negative child each time. Three Rh-negative children in a row is 0.25 cubed, about 1 in 64, which is unusual but entirely ordinary bad luck.
Blood groups and paternity, corrected
The original post had a section headed "Determining Paternity". This one is headed differently, and the difference is the point.
Blood group cannot establish paternity. Not in any case, not partially, not as corroboration.
What it can do is fail to exclude a relationship, which is a much weaker and much more common result.
Published work on exactly this question found that the ABO phenotype could exclude paternity in only about 20 percent of cases. Typing the ABO gene directly does not meaningfully improve on that, because the alleles are simply too common in most populations to be informative very often.
Modern paternity testing uses hundreds of STR DNA markers rather than one gene, and reports a probability of paternity. The result is typically expressed as either 0 or 99.9 percent.
| Method | Can it prove paternity? | How often it excludes |
| ABO blood group | Never | About 20 percent of cases |
| STR DNA testing | Yes, to a stated probability | Nearly all cases |
Blood group can only exclude, and rarely. DNA can include, which is the whole difference.
The original wording was that blood type "can be used to support or rule out biological relationships". The word support is the part that causes harm.
A matching blood group is what you expect between relatives and between unrelated people alike, because four common blood groups are not much information. It is consistent with a relationship and it is also consistent with coincidence, and nothing distinguishes the two.
What ABO typing does not tell you
ABO is one blood group system among many. There are at least 40 recognised, and a patient is commonly typed on several at once, which is why a transfusion request lists more than ABO and Rh.
There is also a version of ABO that standard typing cannot see. In the Bombay phenotype, occurring in under 1 in 1,000 people worldwide, the red cells carry no A, B or H antigen at all.
Such a person appears to be group O on an ABO test. Their plasma, however, contains anti-A, anti-B and anti-H, and the anti-H will destroy O cells that are otherwise perfectly compatible.
This is the clearest reason that blood group alone is never enough and a crossmatch always happens. Type O is not universally safe to receive, and a group O label does not mean the patient is ordinary group O.
The frequencies
| Figure | What it is |
| 1900 to 1901 | when Landsteiner made the discovery |
| 1930 | his Nobel Prize in Physiology or Medicine |
| about 1940 | when the Rh system was defined |
| 4 | ABO blood groups |
| 3 | main ABO alleles, I superscript A, I superscript B and i |
| chromosome 9 | where the ABO gene sits |
| about 20% | of paternity cases the ABO phenotype can exclude |
| 28 weeks | when anti-D prophylaxis is given in pregnancy |
| 72 hours | the window for prophylaxis after delivery |
| under 1 in 1,000 | worldwide frequency of the Bombay phenotype |
| 40 or more | recognised blood group systems in total |
Figures from the sources listed at the foot of this article.
Key Terms in Plain English
These are the words in this article that would send you to a dictionary. Each one is given here the way it is actually used above.
| Term | What it means |
| Antigen | A substance the immune system recognises. A, B and D are blood group antigens. |
| Antibody | A protein that binds a matching antigen. |
| Agglutination | Clumping of red cells when an antibody bridges them. |
| Haemolysis | Bursting of red cells and release of haemoglobin. |
| Crossmatch | The final individual test of donor blood against the patient, before transfusion. |
| Codominant | Both versions of a gene are expressed. This is what makes group AB possible. |
| Sensitisation | Making antibodies after exposure. The process that makes Rh disease possible. |
| Immunoglobulin | An antibody preparation given as a drug. Anti-D immunoglobulin prevents Rh disease. |
| Kernicterus | The neurological injury caused by bilirubin in a newborn. |
| Bombay phenotype | A rare group whose red cells carry no A, B or H antigen. Looks like O on a standard test. |
| Universal donor | Always qualified: O negative for red cells, AB for plasma. Never either type on its own. |
Terms used above, in roughly the order they first appear.
Key Takeaways
- Landsteiner found the ABO system by mixing blood from colleagues, some of whom had died from transfusions that should have worked. He won the 1930 Nobel Prize.
- A blood group follows from one rule: you make antibodies against the antigens you do not have.
- The universal donor for red cells is O NEGATIVE, not O. The universal donor for plasma is AB. The two rules are mirror images.
- Type O is the worst plasma donor there is, because O plasma has both anti-A and anti-B.
- An O negative person can only receive O negative red cells.
- Anti-D is not naturally occurring; you make it only after exposure. Some other anti-Rh antibodies do occur naturally.
- Anti-D prophylaxis at about 28 weeks and within 72 hours of delivery is what prevents Rh disease.
- ABO can also cause hemolytic disease of the newborn, usually more mildly than Rh.
- Blood group cannot prove paternity. It excludes it in only about 20 percent of cases.
- The Bombay phenotype looks like group O on a standard test and is not safe to receive O cells. This is why crossmatch is never skipped.
Frequently Asked Questions
Is O positive a universal donor?
No, and this is the single most common version of the mistake.
An O positive red cell carries the D antigen. Giving one to an Rh-negative person can make their immune system produce anti-D, which then sticks around.
That is harmless once, and a serious problem for a woman who later carries an Rh-positive fetus.
Can blood type prove a child is mine?
No. Blood group can only ever exclude a relationship, never establish one, and it is weak at excluding too.
If a test report says a relationship is "not excluded" on blood grouping, that is not evidence for it. It means the test had nothing to say.
Why do people with the same blood group have different names?
Because a blood group is only part of a full blood profile.
Typing usually covers ABO, Rh, and a set of other systems including Kell, Duffy and Kidd. Two people agreeing on ABO can still differ on all of the others.
Can a person's blood group change?
Not on their own. A blood group is set genetically and stays fixed through life.
There are rare exceptions, chief among them a bone marrow transplant. The person then makes red cells carrying the donor's antigens, and can end up with a different blood group.
Why is my blood group O if both my parents are not O?
Group O needs two copies of the recessive allele, one from each parent.
So both parents must carry it. A type A parent who is IAi and a type B parent who is IBi can, and do, have group O children, as often as one in four.
Is the Bombay phenotype dangerous?
It is dangerous precisely because it looks ordinary. A Bombay patient tests as group O on a standard ABO panel.
Their plasma carries anti-H as well as anti-A and anti-B, and the anti-H destroys ordinary O cells that should have been perfectly safe. Such patients have to be identified before transfusion, not during it.
Key references
• The ABO blood group system
• Blood type and the Rh system
• Karl Landsteiner and the 1930 Nobel Prize
• Why O negative is the universal red cell donor
• Bombay phenotype and why ABO typing is not enough
• ABO blood typing in paternity testing (Bugert et al.)
Images: Wikimedia Commons, with the author and license named in each caption. Figures follow the sources above.
No comments:
Post a Comment