Malcolm Needs

It seems to me that you have covered all bases Mabel. There is a telling sentence in Reid ME, Lomas-Francis C, Olsson ML. The Blood Group Antigen FactsBook. 3rd edition, 2012, Academic Press (page 419), which states, "Experts agree that anti-Yta are often benign and antigen-negative blood may not need to be transfused." Certainly in the cases I have seen over the years, I have never had to give Yt(a-) typed blood (although, because of the geography of England (it's a lot smaller than the USA!), such blood could easily be obtained from the National Frozen Blood Bank, and/or from "tame" donors.

Not only that, but they are amongst the few IgG antibodies that cause agglutination of "normal" red cells without a potentiater (never sure how to spell that - and spelling is my bogey subject anyway!), but also do so at 4oC, through to the warm.

I am sorry BldBnker, but I, amongst others, thoroughly disagree with you about the clinical significance of anti-A1. You must do what you feel safe to do, but I fundamentally disagree. If you look at Marion Reid, Christine Lomas-Francis and Martin Olsson's book, The Blood Group Antigen FactsBook. 3rd edition, 2012, Academic Press, anti-A1 causes either no, or mild/delayed haemolytic transfusion reactions. Their findings are backed up by Geoff Daniels in his book Human Blood Groups.. 3rd Edition, 2013 Wiley-Blackwell, Geoff Daniels and Imelda Bromilow in their book Essential Guide to Blood Groups.. 3rd Edition. 2014, Wiley-Blackwell, Robina Qureshi in her book Introduction to Transfusion Science Practice. 6th Edition. 2015, British Blood Transfusion Society and Harvey Klein and Dave Anstee in their book Mollison’s Blood Transfusion in Clinical Medicine. 12th Edition, 2014, Wiley-Blackwell. In other words, most of the world's leading blood group serologists and blood transfusion doctors disagree. You are perfectly correct when you say that the anti-A1 detected by immediate spin is probably IgM, but, if you look at the findings of haemovigilance organisations throughout the world, most adults have an element of IgG (and IgA come to that) ABO immunoglobulins, and this is particularly so in the case of group O individuals, and, within that cohort, group O females who have been pregnant with an ABO-incompatible foetus. It is these IgG antibodies, in conjunction with the ABO IgM antibodies that cause the worst (often fatal) transfusion reactions, but it is actually the effect of complement that makes these antibodies so clinically significant. In addition, if you perform titration studies on anti-A and anti-B (and anti-A,B), the titres are almost always much higher than the titre of anti-A1, and this, again, influences clinical significance (human-derived high titre anti-A1 is as rare as hen's teeth). The anti-A and/or anti-B in the small amount of plasma left on units that are not ABO identical to the recipient are, therefore, much more likely to cause a transfusion reaction (graft versus host) than is any anti-A1 (host versus graft). Sorry to go on for so long.

Theoretically yes, but in practice no - not unless the need for transfusion becomes more frequent. I have NO idea why three weeks was chosen. There may be a reason, but it seems pretty arbitrary to me!

I am a little worried about the fact that there is no serological cross-match if the mother has made an atypical antibody. The reason I say this is because it is well-known that when a person makes one antibody, they often make more than one. If a mother makes, for example, an anti-K, which is easily detected, she may well also make another antibody specificity, such as an anti-Dia. As the Dia antigen is a low prevalence antigen in most populations, it could well be that the Dia antigen is not expressed on either the screening cells or the antibody identification panel cells - in other words, it may not be detected. Even if the baby does not express the Dia antigen on its red cells, the maternal anti-Dia will still go through the placenta, and so this anti-Dia will still be in the baby's circulation. If, the unit to be transfused is K-, but Di(a+), the baby could well have an unexpected haemolytic transfusion reaction, which could be avoided by a serological cross-match against the mother's sample. Once the unit has been cross-matched, and found to be compatible, then aliquots from the same unit of blood can be safely transfused without a further cross-match, but I feel that, for the first transfusion from any unit of blood, a serological cross-match should be performed.

Sorry Scott, but could I just jump in here? The strength of the reaction of the DAT is not a measure of anything really. It is rather like saying that, if a reaction with anti-D from a pregnant woman is weak, the foetus is not in danger, but I have seen a few cases over the years where the reaction with R1R1 and R2R2 screening cells and panel cells has been weak because the D antigen sites are swamped. The same sort of thing can happen with the DAT.

1. Without a doubt! The thing is that not all antibody specificities will become detectable at exactly the same time (they are a pain, as they do not read the text books or, if they do, they do not take any notice of the bits they do not like)! For example. if you detect an anti-D in the first sample, but nothing else, it does not mean that there is not some other "nasty" bubbling under the level of ability to detect, The next time, there may be, for example, an anti-D and an anti-Jka. It is always better to detect the anti-Jka in vitro, rather than in vivo! 2. In contrast, no, we would not perform an elution every time on such a case, because, in the situation you describe (which, incidentally, working in a Reference Laboratory, we saw only a very regular basis) the patient's immune system is rarely, if ever, working at what may describe as "full capacity". We match them for Rh and K to reduce the chances of "common" antibody production, but that is about all. We would test an eluate about once a month, UNLESS there was clinical evidence of a transfusion reaction (which, given that they are haemolysing their own red cells, may be difficult to detect), or if the time between the need for transfusion becomes noticeably closer, in which case we will do elutions more frequently, but, it should be noted, as there will almost certainly be a panaggltutinin present, the eluate itself may have to undergo alloadsorption, which could weaken the reaction of any underlying atypical alloantibodies.

Just a hunch, but have you tried the Mayo Clinic NY, on the grounds that there is a good possibility that samples would have been sent there for confirmation? These two may not have worked there themselves, but the Mayo may have records of where they worked. You could also try the IBGRL in the UK for the same reasons (possibly samples sent there too, or possibly request for blood through the International Frozen Blood Bank)? These are only vague suggestions.

I haven't heard anything (but then I am retired), but something could come out from the ISBT Meeting taking place in Copenhagen at the moment. I will get my "spies" on to it!

The very FIRST thing I would do is to sign a declaration that any patient who suffers either morbidity or mortality as a result of this change to your system is down to them (i.e. THEY take legal responsibility, and not you). This will have one of two outcomes. Firstly, they will sign such a document, and then you are free from accountability (as long as you record that you have explained to them the error of their ways), or secondly, they will withdraw to a private area to change their under-garments, and will abandon this plan.

I DO understand the business of a method to detect ABO incompatibility Mabel, BUT, if you have a known group A or AB patient with an anti-A1 that, in addition has been shown NOT to be reactive at 37oC, there must, surely, be a way around it? What do you do, for example, if the patient has cold haemagglutination disease, with a high titre, very avid auto-anti-I? Surely, the same must apply, that all units will be apparently ABO incompatible by normal immediate spin techniques? In addition, surely your computer programme should be written so that it helps you, rather than hinders you? It seems to me that, if the human being knows that group A blood, whether it be A1, A2 or any other subgroup of A, will be efficacious for the patient, the computer programme should not prevent the human being from giving that blood, and that the computer programme needs to be rewritten as a matter of urgency. This is particularly important to prevent the wastage of finite group O red cells, but more particularly, if the in vitro findings of Zaffuto et al are found to extend to the in vivo situation (or even if there is the slightest suggestion that they so do).

Well, you can't rely on my memory - it wasn't Transfusion at all, but Vox Sanguinis! The reference is as follows: Zaffuto BJ, Conley GW, Connolly GC, Henrichs KF, Francis CW, Heal JM, Blumberg N, Refaai MA. ABO-immune complex formation and impact on platelet function, red cell structural integrity and haemostasis: an in vitro model of ABO non-identical transfusion. Vox Sanguinis 2016; 110 (3): 219-26.

I will do my very best to look it up over the weekend (between international rugby matches on the television!!!!!!!!!!).

Yes, there are some hospitals in the UK that do not keep group AB red cells. The original excuse/reason was that they did not have that many group AB patients, and so the units expired. As a result, group AB units were, essentially, made "free", in that, if they were not used, the cost was refunded. Still some hospitals did not want to stock them and, privately, I was told this was because they were frightened that they would be given in error to a group A, B or O patient (which is a huge worry, if they think that they and their staff cannot perform accurate ABO typing on patients). In your case, it is different, as you know that you have a "tame" group AB patient who will be, presumably, be using the blood on a regular basis. If you still don't want to transfused group AB blood, I would go for A first, as the anti-B in those units tend to be of lower titre and avidity than does the anti-A in either group B or O units, but, in my opinion, it would be better to give AB red cells. There was a paper recently, I think in Transfusion, but I'm not certain (and I am just about to rush out for an appointment, so I can't check now) that highlighted the fact that ABO antibodies can cause clinical problems other than haemolytic transfusion reactions. Sorry this was written in such great haste.

If the patient is AB, and the anti-A1 does not react at 37oC, I cannot see for a moment why you don't give the patient cross-match compatible AB.