Mabel Adams

I got good advice once from John Judd on an OB with an anti-M reacting in tube with anti-IgG (so it apparently had at least some IgG portion). I couldn't really titrate just the IgG part which is all that really matters in an OB. So he suggested we titrate the "whole" antibody and, as long as the total did not have a titer above 16, we knew the IgG portion alone couldn't. Since many of these anti-M's in pregnancy may not be really significant, this worked great. The titers never rose. We dealt with at least one of these women through more than one pregnancy. We still do our titers in tube because we don't really have a way to correlate the results of gel titers with patient outcomes without collecting data for 30 years (and I don't have the energy right now to figure out how to just correlate the titer results between methods). When we find an antibody in gel, I have been known to run it neat by the titer procedure and if it won't react, I have turned it out as "too weak to titrate." Otherwise we end up turning out titer results of "< 1".

Anything other than anti-D? If just anti-D, it is pretty easy to make sure they are given only to Rh neg pts and a note made that patient received it so any anti-D that shows up later in the patient will be known to be likely to be passive. Other antibodies would be a bigger issue for me.

I finally found a way to describe the technique for getting a bubble. We use the tipmaster pipets. I am not good at it with other pipets. When pipeting the cells, hold the pipet at a 45 deg angle, stabilize it with your other hand, and place the tip as close to the far side of the gel tube opening as you can without touching it. The pressure on the pipet is gentle--just strong enough to complete the stroke. With a little practice it works pretty slick. Hope that helps.

We were taught that it is best to leave an air bubble between the original contents of the gel tube and the cells. I know of another lab that also requires it and they have seen reduced sensitivity when the bubble is not present. Our Ortho trainer did not say it was essential but that it was best (in 2003) to have a bubble. At first, I did not require it, but I found that many techs didn't make much effort to get the bubble and we had more hazy reactions that were hard to interpret. Then I talked to this other lab so now we require it. As I understand it, the logic is that you don't want to mix anti-IgG with your plasma/serum before it has a chance to react with your cells or it could bind up some of the antibody.

We do the same as John. It saves trying to get a sample from a hemorrhaging patient with no BP in a crisis and saves the time of the draw then as well as in more mundane needs without costing much or requiring much work.

For those of you that use anti-A,B on babies/cords, what change in patient care ensues if you find a weak subgroup? Are you comfortable transfusing, say, A units to the baby? Have you found any subgroups only in the anti-A,B test? It could answer an otherwise mysterious case of ABO HDN, I would think. It would register a more accurate type for the baby's records should you ever deal with it as an older patient and it still had the same name etc. (we do cords under the mother's account and don't have an NICU.) Is there any other value doing this test has added to cord/infant testing that I haven't thought of?

I think think part of the 2-3 day rule for doing an antibody screen (i.e. on fresh sample) was based on the deterioration of complement in the sample. If we are using anti-IgG instead of poly, that factor no longer applies. I think IgG is usually pretty stable in a sample, but of course we must follow the package inserts as others have said.

We are dealing with 2 issues here: regulatory and clinical. If you can get someone to state (with a reasonable degree of accuracy) that the patient has been neither transfused nor pregnant in the past 3 months, then you meet the regulatory requirement. For the clinical picture to matter, you would have to have a patient that was recently transfused or pregnant who presents with a neg Ab screen, but who was actually sensitized by the historic circumstance but the antibody titer was too low to detect at the time of the antibody screen. So, what are the odds that when you transfuse them 4-14 days later, the antibody will have acheived a high enough titer to cause a negative clinical outcome? To crunch the numbers (or even get a reasonable estimate) you have to put together the odds that the patient listed as non-pregnant/non-transfused really was; the odds of them being sensitized; the odds of them being in the "window period" of antibody production between the time of the draw and the time of the transfusion, the seriousness of the likely consequences should this all occur and your institution's comfort level with the level of risk this amounts to. What happens if a patient does not accurately report a drug or latex allergy? How excited does the hospital's risk management team get about that? How many times have we all detected "serological transfusion reactions" for which no one detected any signs or symptoms except a need to transfuse again? Should we avoid these at all costs, or is there some cost/benefit ratio to consider? How does a given policy balance the clinical outcome risks and benefit against the resources required to implement the policy>? Hopefully thinking through this will give some perspective. Did I leave anything out?

And a podcast would be advantageous how? Maybe like a teleconference you could take in on your own schedule? Do you actually download a podcast and keep it as a file or is it more like streaming audio that is a one-time deal. If it was something you could keep as a file, it seems like it might have more flexibility (unless the site you get it from is going to keep it available for the 5 years you might want to refer to it).

But you are an icon, Marilyn! I don't think you should be allowed to retire. At least you are willing to share your knowledge online with us, so we appreciate that.

"Mabel, we transfuse around 50 RBC a day, and I cannot think of any transfusion reactions I have seen due to antibodies to low incidence antigens in over 17 years. We pick up 1 or 2 antibodies a year (usually Kpa or Jsa) in our patient who have other antibodies. We do see incompatible units at crossmatch in sickle cell patients who have built a lot of antibodies. Higher incidence of V+ units comes to mind. Over the past 11 years, we have dispensed 60,000+ red cell units using an electronic crossmatch and a 2-cell antibody screen (gel) without a single adverse reaction due to an antibody directed against a low-frequency antigen. I do recall years ago (after the emergence of the type and screen and release of units by immediate-spin crossmatch with a negative antibody screen, reports of patients in two different hospitals who experienced hemolytic transfusion reactions (I believe non-fatal) due to anti-Wra. One hospital decided to return to routine antiglobulin crossmatch and the other did not. Years later, an SBB student at the Los Angeles-Orange Counties Regional Blood Center did a study on litigation involving blood transfusion in the state of California. The results: what little he/she did find regarding serological pretransfusion testing was about ABO incompatibility." Here are 2 answers from my query on the AABB discussion board regarding the frequency of your problem. Hope this helps.

Speaking of odds, how many units of red cells do you transfuse per year?

It's tough to make decisions based on the odds when you have seen the rare consequence occur. Tell your paths to think about all the other things in health care that no one tests the patients in advance for: drug reactions, unexpected bleeding in surgery, pre-existing infections or anemia, etc. etc. (especially nowdays when much less routine labwork is done). Diagnoses almost always play the odds, but they aren't always right the first time. I never hear that any other health care providers feel guilty for these problems.

Based on the odds, you certainly should have your quota of these out of the way for the next few decades. And your nursing staff is clearly paying attention. Can you educate us as to how bad these reactions were?

I have seen many anti-M come up in gel. Cold antibodies do have a characteristic look to them--not in my mind quite like the mixed field reactions Ortho put in the Interpretation Guide but with small amounts of cells spread throughout the column and a bigger button in the bottom than one would expect for the amount of reaction in the column. The only way I know to prove that anti-M is "cold only" is to do a tube test (I do LISS) at the AHG phase only. If it does not react at AHG, I conclude it is "cold only." I think Issitt said that even if anti-M does react at AHG it is only necessary to give xm compatible units, not screened with commercial antiserum. (My experience with anti-M in gel is that there are seldom any weak ones.) I often find the fastest way to find compatible units with a strong anti-M is by doing an IS XM and then doing a gel XM on those that don't react at IS. There is even a CPT code for screening for compatible units using the patient's serum/plasma which I use in these cases.

Did the patient get any sort of immunoglobulin that might have contained anti-C or any donor unit that might have?

After that tragic case that made the papers a few years ago, someone said, "Blood types on transplant donors should be double-checked!" So someone interprets that, without a lot of thought, as doing the test twice. You are right to examine the process for the most likely source of errors and try to plug that hole. "Double-checking" without examining the process is like just throwing a solution at it so you feel better about it, but it doesn't accomplish anything useful. If you are going to expend the resource, make it provide some value. Stick to your guns!

There are 2 different processes here. The traditional rule of 3 relates to having enough data so that the statistical likelihood of a selected antibody identity being due to chance rather than the real antibody is sufficiently low. Thus, we need at least 3 positive and 3 negative cells to react appropriately (or 2 positive and 5 negative) before we can be sure the reactions are not just due to chance. That has nothing to do with what it takes to rule out an antibody. What we really want to do in rule-outs is avoid giving blood that will cause a frank transfusion reaction. We would also like to avoid giving units that will cause delayed or even "serologic" reactions, if we can. Factors that enter into this are: 1. How strong is the antigen I am testing against? 2. How likely are my reagent cells to be damaged in a way to lose sensitivity? 3. How likely is it that the AHG xm won't catch an incompatible unit? 4. How likely is it that the unit will be positive for the antigen (antigen frequency)? 5. How common is this antibody specificity? 6. Is this antibody specificity known for dosage or anamnestic responses? 7. What kind of reaction is most common if a patient reacts to this antigen (intra- or extra-vascular)? 8. How sensitive is my testing method? 9. How easy is it to find a double-dose cell? 10. How easy is it to find a pos cell in the presence of certain other antibodies? I am sure there are more factors that I have overlooked. What I am describing here is actually Failure Mode Effects Analysis for antibody IDs. Wouldn't that make a great student project? All that said, we don't obsess over finding a double-dose (homozygous) Kpa positive cell to rule it out. We seldom even have a positive cell on our screens. We usually balance the risks (severity of consequences) of missing an antibody against the likelihood of that antibody causing a problem. Anti-K is a common antibody but hard to find a double dose cell to test. The odds of transfused units being K pos are relatively low, so if you miss it on a two unit transfusion, odds are that no more than one unit will be incompatible in vivo (and weakly at that or we would have found it). If you miss an anti-Jka, 75% of units will be incompatible, the reaction is more likely to be intravascular, it's a relatively common antibody, it usually isn't too hard to find a double-dose cell to test against and Kidd antibodies are notorious for being anamnestic. Traditional tube testing panels seem to be pretty robust and seldom show much decrease in antigen strength (as long as no one gets near them with bleach), but pre-diluted gel panels seem like they sometimes lose sensitivity (my observation). But gel is a pretty sensitive method. It is tough to rule out anti-E and anti-C in the presence of anti-D with a double dose cell, but any D negative unit is 99+% likely to be negative for C and E as well, so if you miss it the odds are good you will do no harm. Since none of us likes to think this hard every time we do an ID, we usually create some guidelines that we are comfortable with. I don't see any reason for any more than one double-dose rule-out cell unless you have reason to doubt your panel's integrity. For some specificities and in some multiple antibody cases, a single-dose cell is as good as it is going to get. More than one single-dose cell is only helpful to catch mistakes or if one of the cells has a weaker antigen; testing 2 single-dose cells does not magically make the test as sensitive as testing one double-dose cell. There are many labs that don't even do a repeat ID on a patient with a known antibody unless antigen negative units are crossmatch incompatible and they have not seen any particularly high rate of transfusion reactions. We blood bankers probably worry too much about things that have small impact on the patients. Sorry to rattle on so. I guess I should just write an article for Advance or something.

I have thought about only retyping units to detect false positives but haven't done it. By this logic no testing would be required on AB+ units because you would not give them to any patient but an AB+ and if the unit was falsely labeled it would do him no harm. I got an email answer from Ortho regarding their anti-A,B. It is a combination of the same two clones in their anti-A and 2 of the 3 clones in their anti-B. There seems to be no evidence that it reacts more strongly with subgroups. They no longer offer human source anti-A,B. I read the Gammaclone package insert. Their anti-A,B is a mixture of several clones including one that is supposed to react with both A and B antigens (like human source anti-A,. They seem to imply that it reacts more strongly with Ax but don't say anything about its reactions with other subgroups. I have asked for more expert input and will share it if I get some.

Our neg control for IgG is the anti-D test taken through AHG phase (for cells we alternate the A1 and B reverse cells each day because they are known Rh neg); then check cells are added which is the positive control of the IgG. We run a positive plasma from a QC kit and a neg patient control with our pre-diluted gel screening cells. We use our diluent to make up a cell suspension from a unit (that we know will be compatible with our neg control--so usually O unit) and run it with both the neg and pos plasma controls to QC the diluent.

Is human source anti-A,B still readily available? Or did all the manufacturers make a monoclonal of it? I was under the impression that the stuff bottled as monoclonal anti-A & B was just a mixture of monoclonal anti-A and monoclonal Anti-B and it was marketed for retyping O units. If it is a mixture now, is its strength somehow altered so it behaves like human source anti-A,B with weak subgroups? We haven't used it in many years so I have no package insert to read. I know that monoclonal anti-A reacts much more strongly with most subgroups so that the old serological definitions don't really apply to monoclonal anti-A testing. Heaven help the definitions when they finish DNA profiling of them.