[Terr_]

That naming is a little confusing, because if you say "blood group "A" or "B", that indicates the presence of A and B antigen structures, however "blood group H" means a lack of H. (Since almost everybody has H, it was discovered much later.)

Since H is also a building-block needed for A and B to show up, it makes for fun medical-mystery plots, where a child's blood-type seems to be O and doesn't match their parents, but in reality they did get parental A/B/AB genes that just aren't able to express because each parent contributed a nonfunctional H-allele.

https://www.ncbi.nlm.nih.gov/books/NBK2268/

[Terr_]

P.S. IANAPhlebotomist, but if I could wave a wand to change the nomenclature, I'd make it stop relying on an implicit "nothing" state that keeps changing under us as we discover new features that are either usually-present or usually-absent.

Instead patients would have a "blood code" that indicates the tested presence or absence of phenotype cellular features, and any feature not listed would be considered "not yet known, do a test if it might be important."

For example, today's AB- would become +HAB-R for "has H,A,B lacks Rhesus factor." Similarly, O+ would become +HR-AB, and the super rare mutation we were just talking about would be either +R-HAB or -HABR.

Then when we eventually discover a yet-another factor X... Well, yes, your code wouldn't be constant throughout your life, because after an X-test it would gain either a +X or -X... However the upshot is that it eliminates weird implicit guessing games, and medical professionals will "know what they don't know".

[thaumasiotes]

> For example, today's AB- would become +HAB-R for "has H,A,B lacks Rhesus factor."

That's already the way A, B, and Rhesus factor work. The abbreviation for Rhesus factor is "+", but the formal terminology is "Rh+". I would have expected "H-" by analogy.

[Terr_]

> That's already the way A, B, and Rhesus factor work.

No, it isn't: The current ABO system is incapable of expressing "null" data (e.g. "we don't know if B antigens are present or not") and relies on implicit "no marking means false" assumption, which becomes a problem if a new/newly-important antigen feature is discovered, or if a new variation of an existing antigen is found.

For example, suppose we discover a new type Z, which is a membrane-sugar just like A and B, and logically ought to be encoded the same way. It exists in a 50/50 ratio in every existing blood-group, and it finally explains some rare but fatal transfusion problems.

Now some unconscious patient arrives at the emergency room with a wristband saying "AB+". Does that string mean they were tested for Z already, and it was absent, or does it mean there no data about whether Z is present or not? Nobody knows! Worse, if you do a test and discover no Z, then you write the same thing and have the same problem later, all because of the mistake of encoding "false" as an empty string.

Now you might say "Forget the ABO crap, require Z+ or Z- to appear as a separate item", I'll say: "I totally agree, but let's remove the ABO crap and encode all the antigens the same way to make it logically consistent and clear."

> I would have expected "H-" by analogy.

If you're thinking of another section like "A Rh+ H-", then that's close to what I'm proposing, except that it hasn't fixed the historical ambiguities of O/A/AB.

If you fix it to make things explicit and consistent, it becomes "A+ B- Rh+ H- ", and if you shorten Rh to R and group similar things together... that leads to codes like +R-HAB or -ABH+R, which are (ordering aside) what I just proposed.

[Sn0wCoder]

I have never heard of the H-allele. Not sure if @Buttons840 has seen this but “It became a family joke because me and my siblings are all O blood type. If my dad really did have AB blood then we are not his children. The hospital later confirmed it was a mistake.”

Thank you for the information as I always love learning more from the comments than the articles.