Neil Blumberg

While TRALI can be caused by both HLA class I and II antibodies (these antigens are present on most cells (class I) and antigen presenting cells (class II) and HNA (human neutrophil antigens present only on neutrophils in most cases), most respiratory distress after transfusions is caused by other mediators (lipids, CD40L and probably free hemoglobin and heme) and does not require any specific workup of the donor as antibodies are involved only in a few cases (we haven't seen one in years). Some folks in the field do wish to perform antibody studies, but I think they are largely a waste of time, except for transfused or multiparous patients. Others are extremely unlikely to have antibodies in the donor blood. As evidence for these controversial remarks, one need only examine the hemovigilance data from Quebec which found that red cells were primarily involved, which argues against antibodies as a cause in most cases, as there is very little plasma in red cell concentrates. But you might as well follow the standard of practice in your area as to donor workup. I would only send out antibody testing on donors who have a history of transfusion or pregnancy.

I think these bureaucratic methods corrode the trust and collegiality felt by our bedside practice colleagues. High risk patients require discussion between the medical staff of the transfusion service and the ordering provider, and a note in the medical record documenting the decision making. Signing forms is a another tip of the hat to bureaucracy and legalese that should have no role in the provision of medical care. A joint responsibility for such decisions and documentation is far more humane and in the interests of good patient care. Neither the FDA nor regulatory agencies require such divisive practices and we should abandon them for better patient care and documentation of shared decision making. I realize some of you do not have knowledgeable and enthusiastic physician support, but this responsibility needs to be taken by the transfusion service medical director, who is paid to do so, however reluctant and happy to dump this on medical technologists. Just saying.

Just to confuse things a bit, there is absolutely no evidence that fresher units of red cells, or hemoglobin AA units are superior for neonates (or anyone else). The potassium can be an issue, but some fresher units have very high potassiums as well. We define fresh for everyone as 21 days of storage or less. We wash transfusions for newborns in the ICU, or others with high potassium levels, which fixes the potassium issue (I realize this isn't feasible at many hospitals). But I would not worry overly much about storage duration. If concerned, measure the supernatant potassium in a few units and give the one that is lowest. All hospitals can readily measure potassium so why not do that if you are concerned? Storage duration is a poor surrogate for actually measuring potassium. As for CMV seronegative, it's totally unnecessary if the unit is leukoreduced. Not a shred of evidence that it helps and it may preselect for donors who are infected but pre-seroconversion. Save yourself the expense and aggravation and eliminate the use of CMV seronegative units, and convert to 100% leukoreduction if you haven't done that. There are randomized trial data and many observational studies showing the clinical benefits of leukoreduction, including fewer episodes of nosocomial infection, the major cause of hospital morbidity and mortality. Sermon over :).

To much make refractoriness a non-problem, one needs to administer leukoreduced and ABO identical blood components. In our hands, that pretty much abrogates the problem entirely. Almost all of our 3-5 refractory patients per year (out of hundreds) are treated elsewhere first with non-ABO identical components which randomized trial data show increases the rate of HLA alloimmunization. One cannot control what other people do, obviously. So when a patient is clinically refractory, we will HLA type them, try to characterize the antibodies (PRA) and either give HLA matched, or avoid specificities when that is feasible. Either works about as well (probably 70-80% of the time). If those HLA identical platelets are ABO non-identical, they will perform about as badly as non-HLA matched. Thus we try to avoid that or remove incompatible plasma from HLA-matched, plasma incompatible platelets (we wash). It's a good idea to have HPA antibody testing, although not essential, as about 1-2% of refractory patients have HPA antibodies (thus it's rare, but not zero) and this test will also indicate if you have one of the patients who actually have ITP (about 50% of ITP patients have anti-platelet antibodies). There are fairly simple kits available to screen for both HPA and HLA antibodies, although they don't allow for specificity determination.

"Citations are not going to fix this no matter how good an idea it may be." Understood, but if the transfusion medicine leadership, AABB and FDA paid due attention to the data, blood centers would make available products such as plasma reduced group O platelets (whether PAS or washed or otherwise). Or perhaps low-titer group O platelets. If they can do it for whole blood, they could do it for platelets to be used in rural areas and smaller hospitals that cannot inventory a range of products, and cannot wash on site. Blood centers could produce something better than the current products that when used in non-O patients increase bleeding, organ injury and mortality. But apparently, the old incorrect ideas, based upon assumptions that have been proven false, hold sway despite the data. If we ask for safer and more effective products, eventually someone will take the trouble to produce them. Right now, platelet transfusions as currently practiced produce more . harm than benefit. It doesn't help that bedside practitioners have unrealistic expectations about efficacy and act as if this product was some sort of hemostatic glue rather than highly dysfunctional, activated cells that don't work very well and are quite toxic.

"Will PAS platelets with less plasma help?" Yes, the less incompatible antibody (group O mostly) and incompatible soluble and cellular antigen (group A mostly) in the transfusion the better. Most toxicity in medicine is dose dependent, to state what everyone intuitively understands. The ideal platelet product would be washed or otherwise plasma depleted group O platelets. 95% PAS platelets would come close and that's on the horizon. Low titer group O platelets (<50 as for whole blood these days) would be better than what we are doing. We should not let searching for perfection be the enemy of better practices. As a large hospital transfusing 5-6,000 platelet doses (most of which are not indicated ) and with the capability of washing, we almost never give out platelets that aren't ABO identical or group O washed (or group A washed for AB patients) unless the ABO is unknown or it's an emergency. Personally I'd rather wait for 2-12 hours for an ABO identical platelet than transfuse an ineffective and dangerous ABO mismanaged platelet in clinical situations that aren't massive transfusions. The vast majority of our prophylactic transfusions provide no benefit except in acute myeloid leukemia induction therapy. ABO mismatched transfusions actually make bleeding worse in many patients, so why urgently transfuse them to non-bleeding patients? It may cause bleeding, although I don't have that data in prophylactic transfusions. Liberal platelet transfusion certainly caused bleeding in premature newborns who aren't bleeding in the recent randomized trial. I realize our approach simply isn't possible in rural areas and for smaller hospitals. A low titer group O platelet product or 95% PAS group O platelet product would provide greater efficacy and safety without question, but neither is currently available to my knowledge in the USA. It's not rocket science, but for some reason, getting ABO safety right is a low priority for our field. Our patients are the ones who are suffering increased bleeding, organ injury, hemolysis and mortality as the price. The best you can do as a professional faced with these situations is try to get practitioners to defer transfusions for a few hours in non-bleeding patients and not release ABO mismatched platelets without warning the practitioner that this may cause bleeding or make existing bleeding worse. You can cite our work and the work of other groups cited in our review above. You have then discharged your duty and when things improve, as they will eventually, you won't have these unpleasant tasks.

Here are the references. Lancet. 2016 Jun 25;387(10038):2605-2613. doi: 10.1016/S0140-6736(16)30392-0. Epub 2016 May 10. Platelet transfusion versus standard care after acute stroke due to spontaneous cerebral haemorrhage associated with antiplatelet therapy (PATCH): a randomised, open-label, phase 3 trial. Baharoglu MI1, Cordonnier C2, Al-Shahi Salman R3, de Gans K4, Koopman MM5, Brand A5, Majoie CB6, Beenen LF6, Marquering HA7, Vermeulen M1, Nederkoorn PJ1, de Haan RJ8, Roos YB9; PATCH Investigators. Author information Abstract BACKGROUND: Platelet transfusion after acute spontaneous primary intracerebral haemorrhage in people taking antiplatelet therapy might reduce death or dependence by reducing the extent of the haemorrhage. We aimed to investigate whether platelet transfusionwith standard care, compared with standard care alone, reduced death or dependence after intracerebral haemorrhage associated with antiplatelet therapy use. METHODS: We did this multicentre, open-label, masked-endpoint, randomised trial at 60 hospitals in the Netherlands, UK, and France. We enrolled adults within 6 h of supratentorial intracerebral haemorrhage symptom onset if they had used antiplatelet therapy for at least 7 days beforehand and had a Glasgow Coma Scale score of at least 8. With use of a secure web-based system that concealed allocation and used biased coin randomisation, study collaborators randomly assigned participants (1:1; stratified by hospital and type of antiplatelet therapy) to receive either standard care or standard care with platelet transfusion within 90 min of diagnostic brain imaging. Participants and local investigators giving interventions were not masked to treatment allocation, but allocation was concealed from outcome assessors and investigators analysing data. The primary outcome was shift towards death or dependence rated on the modified Rankin Scale (mRS) at 3 months, and analysed by ordinal logistic regression, adjusted for stratification variables and the Intracerebral Haemorrhage Score. The primary analysis was done in the intention-to-treat population and safety analyses were done in the intention-to-treat and as-treated populations. This trial is registered with the Netherlands Trial Register, number NTR1303, and is now closed. FINDINGS: Between Feb 4, 2009, and Oct 8, 2015, 41 sites enrolled 190 participants. 97 participants were randomly assigned to platelet transfusion and 93 to standard care. The odds of death or dependence at 3 months were higher in the platelet transfusiongroup than in the standard care group (adjusted common odds ratio 2·05, 95% CI 1·18-3·56; p=0·0114). 40 (42%) participants who received platelet transfusion had a serious adverse event during their hospital stay, as did 28 (29%) who received standard care. 23 (24%) participants assigned to platelet transfusion and 16 (17%) assigned to standard care died during hospital stay. INTERPRETATION: Platelet transfusion seems inferior to standard care for people taking antiplatelet therapy before intracerebral haemorrhage. Platelet transfusion cannot be recommended for this indication in clinical practice. N Engl J Med. 2019 Jan 17;380(3):242-251. doi: 10.1056/NEJMoa1807320. Epub 2018 Nov 2. Randomized Trial of Platelet-Transfusion Thresholds in Neonates. Curley A1, Stanworth SJ1, Willoughby K1, Fustolo-Gunnink SF1, Venkatesh V1, Hudson C1, Deary A1, Hodge R1, Hopkins V1, Lopez Santamaria B1, Mora A1, Llewelyn C1, D'Amore A1, Khan R1, Onland W1, Lopriore E1, Fijnvandraat K1, New H1, Clarke P1, Watts T1; PlaNeT2 MATISSE Collaborators. Collaborators (124) Author information Abstract BACKGROUND: Platelet transfusions are commonly used to prevent bleeding in preterm infants with thrombocytopenia. Data are lacking to provide guidance regarding thresholds for prophylactic platelet transfusions in preterm neonates with severe thrombocytopenia. METHODS: In this multicenter trial, we randomly assigned infants born at less than 34 weeks of gestation in whom severe thrombocytopenia developed to receive a platelet transfusion at platelet-count thresholds of 50,000 per cubic millimeter (high-threshold group) or 25,000 per cubic millimeter (low-threshold group). Bleeding was documented prospectively with the use of a validated bleeding-assessment tool. The primary outcome was death or new major bleeding within 28 days after randomization. RESULTS: A total of 660 infants (median birth weight, 740 g; and median gestational age, 26.6 weeks) underwent randomization. In the high-threshold group, 90% of the infants (296 of 328 infants) received at least one platelet transfusion, as compared with 53% (177 of 331 infants) in the low-threshold group. A new major bleeding episode or death occurred in 26% of the infants (85 of 324) in the high-threshold group and in 19% (61 of 329) in the low-threshold group (odds ratio, 1.57; 95% confidence interval [CI], 1.06 to 2.32; P=0.02). There was no significant difference between the groups with respect to rates of serious adverse events (25% in the high-threshold group and 22% in the low-threshold group; odds ratio, 1.14; 95% CI, 0.78 to 1.67). CONCLUSIONS: Among preterm infants with severe thrombocytopenia, those randomly assigned to receive platelet transfusions at a platelet-count threshold of 50,000 per cubic millimeter had a significantly higher rate of death or major bleeding within 28 days after randomization than those who received platelet transfusions at a platelet-count threshold of 25,000 per cubic millimeter. (Funded by the National Health Service Blood and Transplant Research and Development Committee and others; Current Controlled Trials number, ISRCTN87736839 .). Front Immunol. 2015 Feb 2;6:28. doi: 10.3389/fimmu.2015.00028. eCollection 2015. Platelet transfusion - the new immunology of an old therapy. Stolla M1, Refaai MA1, Heal JM1, Spinelli SL1, Garraud O2, Phipps RP3, Blumberg N1. Author information Abstract Platelet transfusion has been a vital therapeutic approach in patients with hematologic malignancies for close to half a century. Randomized trials show that prophylactic platelet transfusions mitigate bleeding in patients with acute myeloid leukemia. However, even with prophylactic transfusions, as many as 75% of patients, experience hemorrhage. While platelet transfusion efficacy is modest, questions and concerns have arisen about the risks of platelet transfusion therapy. The acknowledged serious risks of platelet transfusion include viral transmission, bacterial sepsis, and acute lung injury. Less serious adverse effects include allergic and non-hemolytic febrile reactions. Rare hemolytic reactions have occurred due to a common policy of transfusing without regard to ABO type. In the last decade or so, new concerns have arisen; platelet-derived lipids are implicated in transfusion-related acute lung injury after transfusion. With the recognition that platelets are immune cells came the discoveries that supernatant IL-6, IL-27 sCD40L, and OX40L are closely linked to febrile reactions and sCD40L with acute lung injury. Platelet transfusions are pro-inflammatory, and may be pro-thrombotic. Anti-A and anti-B can bind to incompatible recipient or donor platelets and soluble antigens, impair hemostasis and thus increase bleeding. Finally, stored platelet supernatants contain biological mediators such as VEGF and TGF-β1 that may compromise the host versus tumor response. This is particularly of concern in patients receiving many platelet transfusions, as for acute leukemia. New evidence suggests that removing stored supernatant will improve clinical outcomes. This new view of platelets as pro-inflammatory and immunomodulatory agents suggests that innovative approaches to improving platelet storage and pre-transfusion manipulations to reduce toxicity could substantially improve the efficacy and safety of this long-employed therapy.

My main concern is that there are virtually no clinical situations where two doses of platelets are indicated. Platelets are largely ineffective and dangerous in bleeding patients, and doubling the dose, doubles the risk but does nothing to increase efficacy. These products are not "glue" they are highly activated, abnormal inflammatory and pro-thrombotic cells. At least that is what the literature and new randomized trials show. Platelet transfusion as now practiced increases bleeding by largely unknown mechanisms in stroke patients and newborns who are aggressively transfused. This double dosing is simply a practice that should be abandoned as ineffective and dangerous.

Randomized trials from the 1990s show that clinical refractoriness (a death sentence pretty much) is two times higher when randomly selected ABO platelets are transfused (e.g., whatever is about to outdate) and five times higher when ABO incompatible are intentionally transfused. The rest of the data come from laboratory studies and observational clinical data. ABO non-identical platelets are associated with 50% increases in red cell use in trauma patients and and 15% higher mortality rates. In cardiac surgery, the receipt of two or more ABO non-identical platelets increases red cell transfusions by 50% and mortality by several fold compared with patients matched for the same number of ABO identical platelets. ABO non-identical plasma is associated with significant increases in sepsis and acute lung injury in surgical patients, after adjustment for other prognostic factors. In vitro, ABO immune complexes cause hemolysis in group O red cells, interfere with platelet and endothelial cell function. If it walks like a duck and quacks like a duck, it's likely a duck :). We have all the knowledge we need to decide to stop prioritizing inventory management and prioritize not infusing incompatible ABO antigen and antibody, and it is feasible if you put your back into it in a larger hospital. Harder for AB patients and smaller hospitals, but we can at least stop pretending that universal donor products are equivalent to ABO identical products in efficacy and safety. They are not, even if we use them in emergencies. The references for all this (about a dozen studies) are in the Anesthesia and Analgesia paper listed above if anyone is interested. The other issue is that those that reject these conclusions have no data of their own to show that ABO non-identical platelets and other products are effective and safe, only expert opinion and historical habits of practice.

What are you referring to with "95% of recipients of AB plasma have antibody to the soluble antigen present"? I'm referring to the unknown ABO type patient who receives AB plasma, or the intentional use of AB plasma for all patients despite knowing the recipient's ABO type. All of those O, A and B patients have antibody that will react with the A and B soluble antigens that are present in AB plasma (even in non-secretors, albeit in smaller amounts). We and others have empirical and laboratory evidence that these immune complexes interfere with platelet and endothelial cell function/integrity leading to increased bleeding, lung injury, sepsis and mortality. Even with low titer O Pos WB units aren't you going to see just as much formation of antigen-antibody complexes as you do with a non-type specific platvelet pheresis? The volume of plasma will be about the same and will contain anti-A, anti-B and anti-A,B, correct? Don't know for sure, but a little bit of antibody in a body full of antigen may or may not be as dangerous as a little bit of antigen in the presence of lots of antibody (think about anaphylaxis versus the safety of IVIgG). The biologic and clinical consequences of immune complex formation may well be dependent on ratios of antigen to antibody, and the size of the complexes formed. In general, lots of antibody on a bit of antigen seems to more biologically active than a lot of antigen with modest amounts of antibody. What about the Trauma centers that are using A plasma - is that any better? One of the first clues is that A plasma is no worse than AB, suggesting that the antibody content makes little difference. Group A plasma is at least identical with 40% of recipients. Time will tell, but I suspect that low titer O plasma and whole blood will be no worse than our current practices where we give tons of incompatible soluble antigen (and sometimes cellular antigen with platelets).

No need for platelets if the whole blood is stored less than about 14-15 days. Room temperature stored platelets aren't very effective in any case, particularly if ABO mismatched, and refrigerated whole blood platelets still retain some function for considerably longer than previously thought. The main attraction for advocates is ease of use, as well as transportation simplicity to the site of traumatic injury. There are those who believe that cold stored platelets will be more effective than the customary room temperature stored platelets. There is no denying the first two points, but the third is still uncertain as far as clinical efficacy and safety. That said, we've seriously overestimated the efficacy and seriously underestimated the toxicity of current platelet products (particularly those that are not ABO identical). Two recent randomized trials show that liberal platelet transfusion worsens bleeding and increases mortality compared with no transfusion or restrictive platelet transfusion. Lancet. 2016 Jun 25;387(10038):2605-2613. doi: 10.1016/S0140-6736(16)30392-0. N Engl J Med. 2019 Jan 17;380(3):242-251. doi: 10.1056/NEJMoa1807320. In addition to logistics, one major challenge is going to be the use of so-called universal donor O whole blood in the 55% of the population that is not group O. If the whole blood is really low-titer (<50 or so) and only a few units are transfused, this practice is probably not likely to be a big problem for patients. But we may see "indication creep," and use of group O whole blood (but not so low titer, <200 :)) ) both in non-hospital situations and then in hospital treatment of patients of known ABO type (not O). This will potentially cause low grade hemolysis, bleeding, multi-organ failure and mortality if we are not careful about finding out what is both safe and effective (see references below). Fortunately for the advocates of whole blood, and hopefully, for our patients, the current practices of so-called "universal donor"group O red cells, whatever ABO type platelets and so-called "universal donor" AB plasma for all patients is highly toxic due to formation of ABO immune complexes in every patient we give this cocktail to. Leads to hemolysis, endothelial cell injury, organ failure and increased bleeding in the patients we have reported, as compared with patients receiving only ABO identical transfusions. Group O whole blood is ABO identical with 45% of recipients, which will be progress from 0% with our current protocols since 95% of recipients of AB plasma have antibody to the soluble antigen present. My predicition is that If the amount of incompatible group O plasma from whole blood transfused is modest, and the titer of anti-A/B is low enough, whole blood may indeed be safer than what we are currently transfusing. Time will tell. Vox Sang. 2011 Jul;101(1):55-60. doi: 10.1111/j.1423-0410.2010.01464.x. Anesth Analg. 2018 Jun;126(6):2135-2138. doi: 10.1213/ANE.0000000000002600.

If you have access to a maternity service, cord cells are negative for CD38, thus we use a homemade cord cell panel, which has been inexpensive and 100% successful in screening for clinically significant anti-red cell antibodies. This drug is going to be used more frequently in initial treatment for multiple myeloma so one might as well be prepared for doing this very frequently in the future. Fortunately these patients are not heavily transfused and only occasionally have alloantibodies. Transfusion. 2015 Sep;55(9):2292-3. doi: 10.1111/trf.13174. Alternatively, you can use DTT, see the above letter for references. Daratumumab cord cell method.pdf

Another expensive but possibly effective approach would be to use one of the complement activation inhibitors, such as eculizamab, that is used to treat paroxysmal nocturnal hemoglobinuria, a disease with inadequate inactivation of complement components.

Another strategy, which works for ABO incompatible kidney transplants in some cases, is a combination of immunosuppressive drug therapy, IVIgG and plasma exchange. If it works for ABO, one would guess that it could work for Inb (or anything else, for that matter). One also guesses that the antibody might be wholly or largely IgM if it only causes HTR and not HDN. If that were the case, plasma exchange could be particularly effective.

I agree that blood bank refrigerators in the OR (or anywhere else) are accidents waiting for a place to happen. We use a temperature monitored and controlled cooler system so that the blood for the patient in the OR is sitting right next to the anesthesiologist. Have had no mistransfusion accidents in the OR in the close to 40 years I've been here. In the hospital where I trained there was an OR refrigerator and we had mistransfusions every few months. Case closed.

Nurses are supposed to know how to give intramuscular injections to heavier patients. That said, it can be given intravenously or sub-cutaneously with equal benefit, although few outpatient practices give intravenous injections. With very heavy patients, a longer needle works.

Most transfusion practices in neonatalogy, including this one, are not evidence based, but rather empirical expert opinion. The use of reconstituted whole blood is more historical than anything else. A unit or two of recently collected (perhaps 7-14 days) whole blood would probably be as rational. One might check the potassium before using to make sure it isn't super high. That is the rationale for washing a red cell. It removes potassium from hemolysis during collection and storage, and makes the red cells more likely to absorb potassium once transfused. It's definitely more useful if the baby is hyperkalemic to begin with. Otherwise, whole blood would be fine. The reason for using plasma is fear of hypocoagulability, which is probably mostly mumbo jumbo for small exchanges, but might be more of an issue for larger exchanges (2 or more blood volumes). There is no real proof that any of these approaches is superior or inferior. Calculating the hematocrit is a case of weighing the red cells and measuring their hematocrit and then diluting accordingly with plasma or albumin solution (5%). You don't want a hematocrit higher than 40 in the exchange as normal neonates do not have high hematocrits and oxygen delivery is actually worse at hematocrits much above 30 in experimental models. In this case, more is not better as far as anyone knows. Once again, this is expert opinion not evidence based.

If you have access to a medical school or teaching hospital library, it will likely be available there at one of their computer stations.

One might add that use of transfusion in acute chest syndrome has almost no evidence base and may indeed do more harm than good. In transfusion, less is usually better than more.

" I know we will be getting another order for an Erythrocytapheresis" That's not to say that is good clinical judgment. There is not a shred of evidence that 30% is better than 35%, and to expose the patient to the risk of another 6-8 units of red cells and an invasive procedure is, in my opinion a serious misreading of the benefits and risks of transfusion in this setting. Not to mention that the same goal (30%) could be achieved by simple transfusion of AA red cells. But neither apheresis exchange nor transfusion is indicated for clinical benefit.

One unit of hemoglobin AS blood will contribute perhaps 5% to the S level in a patient with hemoglobin SS whose overall post-apheresis hemoglobin level is 10 g/dl. Not enough to make a real difference. A lot of work for pretty much nothing in my view. But we do it anyway. Probably would be less work to test donor units if the %S doesn't drop quite as far as expected. Hemoglobin S in AS red cells does not behave the way it does in SS red cells, so this has absolutely no clinical implications. The hemoglobin S goal for most treatment is 30% or less. 5% is not going to make any difference in treatment plans. Just an alternative view from the usual :).

There is no such thing as never in science and medicine. But while leukoreduced transfusions may on rare occasions be associated with a CMV seroconversion, the same is true of CMV seronegative, since it is possible to have a donor who is viremic but not yet seropositive. There are those who believe CMV is almost never transmitted by transfusion, but that these seroconversions are by the usual route of individual to individual environmental transmission. I am close to that point of view. We have not used CMV seronegative, as pointed out above, for the last 20 years plus. We have a 70 bed+ neonatal intensive care unit, do about 80-100 allogeneic transplants of stem cells, heart transplants, etc. CMV seronegative is totally unnecessary and provides little or no benefit to patients. Leukoreduction is much more important overall and provides enough CMV safety on its own, in my view, to beat a dead horse here :).

This is something that only works when there is expert physician to physician communication. Your medical director needs to undertake this project. There are substantial data from randomized trials and observational cohort studies that leukoreduction abrogates CMV seroconversion. These are the studies we used twenty plus years ago to convince our practitioners that leukoreduction was not only good enough, but almost certainly superior in overall clinical outcomes to CMV seronegative non-leukoreduced transfusions. Of course no patient should be receiving non-leukoreduced transfusions at this late date, but in the USA not all transfusion medicine physicians are convinced of this, in my opinion, strongly justified clinical practice.

We do the same thing, for historical reasons. Probably totally unnecessary :). The amount of S hemoglobin in one unit of donor blood is not clinically significant in patient management.

No guidelines, just clinical common sense. There are absolutely no data to support the need for excluding heterozygous donors, nor the need for heterozygous recipients to receive only AA blood. Heterozygous patients are physiologically normal except for some data to suggest that under extreme conditions of dehydration, altitude they are slightly more susceptible to complications that occur even among hemoglobin AA patients.