Dear Colleagues,

I would like to get your comments on the topic.

We are a tertiary care health set up in Western India and want to start ab screening for all patients who might require blood transfusion. Just to appraise you of situation in our part of the world regarding Blood Banking:

1. Antigen profile of the populatin is not available yet. Whatever studies have been done are not true representation of the entire population.

2. Only fewcentres are aware of (still cannot say well versed) the immunohematology techniques applicable and still few practice them. This to the extent, that even tube technique for cross matching (with or without AHG) is not know to many blood banks.

We group & type our patients for ABO and Rh(D) and perform gel crossmatch. Any ab found is resolved 'retrospectively' .

Now, coming to the query - Studies done in India (2-3) have shown that sensitivity of (picking ab by) currently available cell panel is only 80% i.e. no ab on cell panel still crossmatch incompatible. These were for clinically significant abs.

Q1) So what is the solution? How do we go about it?

Q2) what are the advantages of ab screening in pre-transfusion sample in this scenario?

Thank you

You are saying that 20% antibodies are not being detected on your screen/ panel but are being picked up on crossmatch ?

This is extremely high. Have you eventually identified these antibodies- and if so what were they (you mentioned they were clinically significant).

In the UK we have a fairly diverse population, and have used the National Blood Service manufactured screens and panels (which I presume are mainly from caucasian donors)- had no noticable problems with transfusions in mixed recipient populations.

I must admit, I am struggling to think of any antigens found on the Indian sub-continent that would cause such a high frequency of false negative reactions. Even within this population, the In(a) antigen is not that common.

I know that in the Peoples' Republic of China they have to include red cells expressing Mi(a) and Di(a) because of the relatively high frequency of both these antibodies and antigens in their population, but I still can't think of such a combination in India.

I worry that it is the technique being used, or that cells expressing single dose of antigens (e.g. S+s+, Fy(a+b+), Jk(a+b+) are being used in the screen - screening cells should always express single dose) or that the red cells are not kept in a preservative/stored at a sub-optimal temperature (or a combination) but other than that I am stumped.

A few quick intial thoughts for you.

1) Are you doing a DAT on the donor. If the donor has a positive DAT, an AHG XM will be incompatible.

2) If the patient has Bg antibodies, a unit might be incompatible, but the panel may be negative.

3) If an ABO antibody is present (i.e. Anti-A1 in A2 patient), the unit may be incompatible and the panel negative.

4) The unit is P1 strong, the patient has a weak Anti-P1 and none of the panel cells are P1 strong.

5) The patient has an Anti-Sd(a) and the unit is Cad+, etc.

Of course, an antibody to a low incidence antigen remains a possibility, but it's not the only one.

A few quick intial thoughts for you.

1) Are you doing a DAT on the donor. If the donor has a positive DAT, an AHG XM will be incompatible.

2) If the patient has Bg antibodies, a unit might be incompatible, but the panel may be negative.

3) If an ABO antibody is present (i.e. Anti-A1 in A2 patient), the unit may be incompatible and the panel negative.

4) The unit is P1 strong, the patient has a weak Anti-P1 and none of the panel cells are P1 strong.

5) The patient has an Anti-Sd(a) and the unit is Cad+, etc.

Of course, an antibody to a low incidence antigen remains a possibility, but it's not the only one.

I agree with all you are saying, but even then, I cannot see it adding up to a 20% discrepancy.

You are saying that 20% antibodies are not being detected on your screen/ panel but are being picked up on crossmatch ?

This is extremely high. Have you eventually identified these antibodies- and if so what were they (you mentioned they were clinically significant).

In the UK we have a fairly diverse population, and have used the National Blood Service manufactured screens and panels (which I presume are mainly from caucasian donors)- had no noticable problems with transfusions in mixed recipient populations.

Thanks for the reply.

We are still a new set up and what I quoted was from 2 good centres in India (both are teaching institute with transfusion courses running there). Author of the study only said, "missed abs were clinically significant", without mentioning the specificity.

We are planning to start the ab screening for all our IP patients and this was recently OKeyed by our Hospital Transfusion Committee. Thus this query of mine.

Do you routinely crossmatch (AHG) patient samples negative on ab screen? Or is it just IS/ Electronic Xm?

Thanks for the reply.

You are correct. We still don't have antigen profile available for our population. Big studies are underway for the same. This will help companies develop indigenous cell panel. Also, study (done at one big centre in India, and not by us) did find one anti-Mi(a) in the patient population.

We are beginning this exercise with our population. Till now we were doing Grouping and crossmatching without ab screen for our patients.

Actually, transfusionpgi, I wouldn't be at all worried that you found an anti-Mia in your population. It's not the frequency of the antibody that matters, but reather the frequency of the antigen in the population.

We recently had an Lu:14 in one of our panels (I know not why) and, as a consequence we found a couple of anti-Lu14s, but they didn't matter because Lu14 is so rare.

Also, you will find that, very often, once an individual has made an antibody against one low frequency antigen, they very often make a whole soup of antibodies against other low frequency antigens. You also have to be mighty careful that the red cells you are using that are "typed" as, for example, Wr(a+) are really Wr(a+). In other words, they have been typed using a reputable monospecific anti-Wra (or genotyped) and not with an antiserum labelled as anti-Wra, but which itself may have many other antibodies in it directed against low incidence antigens.

It is an area fraught with danger for the unwary!

Thanks for the reply.

We are still a new set up and what I quoted was from 2 good centres in India (both are teaching institute with transfusion courses running there). Author of the study only said, "missed abs were clinically significant", without mentioning the specificity.

We are planning to start the ab screening for all our IP patients and this was recently OKeyed by our Hospital Transfusion Committee. Thus this query of mine.

Do you routinely crossmatch (AHG) patient samples negative on ab screen? Or is it just IS/ Electronic Xm?

We electronic issue - if screen neg and no previous history of antibodies, also require 2 x groups on LIMS.

I would have thought that until you resolve your 20% incompatibility problem, you would need to carry on with performing this.

How reliable is the testing (ABO, screening, etc) performed on your donor units...do you find many incorrectly labelled ones, or does the blood service have good donor unit tracking, from the point of collection to final bag labelling ?

It's very interesting to learn of the challenges facing blood bank folk in other countries.

Hi There from a new member. I am the Technical Director of an Australian Immunohaematology manufacturer but I spend most of my time in Asia. I often see the picture you describe. As an example, in the Philippines, some Hospitals report that around 30% of antibodies found in crossmatches are not detectable on caucasian screening cells and do not react with caucasian ID panels. They are not auto or so called false negatives. They are really alloantibodies that have specificity for antigens that are rare in caucasian populations but are relatively common in Asian populations. These antibodies in this population are likely to be vMNS (the new nomenclature for the 11 antigens formerly known as Miltenberger or Mi and includes MUT, Mur and Mia and others). Some will also be Diego a that is not commonly found on screening panels.

A literature search shows that these antigens and antibodies are not characterised in India itself but published studies (Prathiba, 2002) in Malaysian Indian populations show that 3% of that population are vMNS or Mia positive. Malaysian Chinese are 4.9% and Malay 2.8%. vMNS antibodies have also been found in the Sri Lankan population whose dominant Sinhalese population are ethnically distinct from but have similar phenotypes to the Indian sub-continental populations.

As an example a very large study in 2003 by Lee et al showed that the the alloantibody incidence in Hong Kong antenatal patients was 0.27% and 57.6% of these were classed as Anti-Mi. The antigen incidence of Mia in HK was shown to be 6.3 by Poole et al in 1991. This antigen incidence is right in the ball park to generate a very significant antibody incidence. Lots of antigen negative blood recipients and mothers but a significant antigen positive number of blood donors and babies.

I would point out a few things about these studies.

1. They were state of the art at the time but the knowlege about vMNS has changed in the last few years. The antibody screening was done with natural cells that were shown to be Mi positive by the only monoclonal antibody available at the time. In reality they are uncharacterised and are likely to be what we would now call Mi III. In this case, they may miss some vMNS antibodies, so the real incidence is likely to actually be higher.

2. vMNS antibodies seem to behave like other antibodies of the MNS system in that there are so called "naturally occuring" forms that are probably better called "non-red cell immune" and these are IgM and cold reacting and likely to be clinically irrelevant. Others are alloimmune in nnature, IgG and clearly very capable of causing severe transfusion reactions and HDFN. Some as yet unpublished work has shown that in some populations that up to 40% of vMNS antibodies are IgM and likely of little clinical harm. The other 60% we should be very worried about.

For this reason, I hold the opinion that testing laboratories in Asian countries have to continue to IAT crossmatch to detect these antibodies. I can back this up with many anecdotal cases where I have seen Asian laboratories adopt western style group and screen blood release protocols only to find a very significant number of transfusion reaction in patients who are antibody screen negative (using imported western panels) and who are transfused ABO compatible blood. If you have a look at Taiwan as an example. They have published a large body of work and are very well aware of the high incidence of vMNS antibodies and the clinical relevance of them in their population. They therefore tend to screen with an imported panel BUT also add a locally made 3 cell panel that would be unusable in a western country but does have Mia and Dia on it. So they routinely do a 6 cell antibody screen to try to cover all the antibodies in their population. They also tend to use the Polybrene technique that I personally do not like but they feel has high sensitiivity to vMNS antibodies.

To date, India has not been well studied and certainly nothing substantial has been published. I think that while the Indian phgenotypes are significantly different from other more Eastern asian populations, there is sufficient evidence to beleive that the Indian population will have a very significant incidence of very clinically relevant vMNS antibodies that will not be detected on normal imported panels.

So transfusionpgi, as a final note, my group has very recently come up with a solution to this problem. We have and are running a number of studies in Asian countries to update or define the incidence of red cell antigens and the incidence and specificity of alloantibodies. PM me if you would like to know more or are interested in participating in a study I am planning.

A very fine answer TimOZ.

I agree with everything you say.

Dear Transfusion gpi

I see you are doing a gel crossmatch. I presume the gel is DiaMed gel. Am I correct in thinking that you are also using DiaMed screening cells to do the antibody screen? Or are you just sending the sample off for an identification panel if the crossmatch is positive? Either way, you should really get in touch with DiaMed India in Gurgaon (always assuming you are using diamed gel). I can promise you they have an EXCELLENT team of specialists who will be able to advise you.

Dear Transfusion gpi

I see you are doing a gel crossmatch. I presume the gel is DiaMed gel. Am I correct in thinking that you are also using DiaMed screening cells to do the antibody screen? Or are you just sending the sample off for an identification panel if the crossmatch is positive? Either way, you should really get in touch with DiaMed India in Gurgaon (always assuming you are using diamed gel). I can promise you they have an EXCELLENT team of specialists who will be able to advise you.

Dear Galvania!

You are bang on target! We indeed use Diamed gel cards and send anything which cannot be resolved at our end to Diamed people. How do you know that they have their headquarter in Gurgaon?

Anyway, since Diamed is one of the stake holder in the entire seen, their view is bound to be biased and in fact is biased, as they do not even recommend Aisan Diacell panel (which I presume is short in supply and thus they don't mind dispatching whatever is available).

You seem to be related to Diamed in someway, is it?

We electronic issue - if screen neg and no previous history of antibodies, also require 2 x groups on LIMS.

I would have thought that until you resolve your 20% incompatibility problem, you would need to carry on with performing this.

How reliable is the testing (ABO, screening, etc) performed on your donor units...do you find many incorrectly labelled ones, or does the blood service have good donor unit tracking, from the point of collection to final bag labelling ?

It's very interesting to learn of the challenges facing blood bank folk in other countries.

Electronic issue is still a 'far cry' for us. I am sure your LIMS must be validated nicely! Are you following some guidelines?

We plan to continue it (AHG Xm) till we either have our cell panel or population profile done (for ags).

Now, there again is a 'news' for you! Blood supply here doesn't have segregation as blood suppliers and transfusion centres/ hospitals (more of it some other time), but ...Yes, labeling and other processes are well validated and we don't find these kind of errors anymore (yes, we did have them initially though)

Hi There from a new member. I am the Technical Director of an Australian Immunohaematology manufacturer but I spend most of my time in Asia. I often see the picture you describe. As an example, in the Philippines, some Hospitals report that around 30% of antibodies found in crossmatches are not detectable on caucasian screening cells and do not react with caucasian ID panels. They are not auto or so called false negatives. They are really alloantibodies that have specificity for antigens that are rare in caucasian populations but are relatively common in Asian populations. These antibodies in this population are likely to be vMNS (the new nomenclature for the 11 antigens formerly known as Miltenberger or Mi and includes MUT, Mur and Mia and others). Some will also be Diego a that is not commonly found on screening panels.

A literature search shows that these antigens and antibodies are not characterised in India itself but published studies (Prathiba, 2002) in Malaysian Indian populations show that 3% of that population are vMNS or Mia positive. Malaysian Chinese are 4.9% and Malay 2.8%. vMNS antibodies have also been found in the Sri Lankan population whose dominant Sinhalese population are ethnically distinct from but have similar phenotypes to the Indian sub-continental populations.

As an example a very large study in 2003 by Lee et al showed that the the alloantibody incidence in Hong Kong antenatal patients was 0.27% and 57.6% of these were classed as Anti-Mi. The antigen incidence of Mia in HK was shown to be 6.3 by Poole et al in 1991. This antigen incidence is right in the ball park to generate a very significant antibody incidence. Lots of antigen negative blood recipients and mothers but a significant antigen positive number of blood donors and babies.

I would point out a few things about these studies.

1. They were state of the art at the time but the knowlege about vMNS has changed in the last few years. The antibody screening was done with natural cells that were shown to be Mi positive by the only monoclonal antibody available at the time. In reality they are uncharacterised and are likely to be what we would now call Mi III. In this case, they may miss some vMNS antibodies, so the real incidence is likely to actually be higher.

2. vMNS antibodies seem to behave like other antibodies of the MNS system in that there are so called "naturally occuring" forms that are probably better called "non-red cell immune" and these are IgM and cold reacting and likely to be clinically irrelevant. Others are alloimmune in nnature, IgG and clearly very capable of causing severe transfusion reactions and HDFN. Some as yet unpublished work has shown that in some populations that up to 40% of vMNS antibodies are IgM and likely of little clinical harm. The other 60% we should be very worried about.

For this reason, I hold the opinion that testing laboratories in Asian countries have to continue to IAT crossmatch to detect these antibodies. I can back this up with many anecdotal cases where I have seen Asian laboratories adopt western style group and screen blood release protocols only to find a very significant number of transfusion reaction in patients who are antibody screen negative (using imported western panels) and who are transfused ABO compatible blood. If you have a look at Taiwan as an example. They have published a large body of work and are very well aware of the high incidence of vMNS antibodies and the clinical relevance of them in their population. They therefore tend to screen with an imported panel BUT also add a locally made 3 cell panel that would be unusable in a western country but does have Mia and Dia on it. So they routinely do a 6 cell antibody screen to try to cover all the antibodies in their population. They also tend to use the Polybrene technique that I personally do not like but they feel has high sensitiivity to vMNS antibodies.

To date, India has not been well studied and certainly nothing substantial has been published. I think that while the Indian phgenotypes are significantly different from other more Eastern asian populations, there is sufficient evidence to beleive that the Indian population will have a very significant incidence of very clinically relevant vMNS antibodies that will not be detected on normal imported panels.

So transfusionpgi, as a final note, my group has very recently come up with a solution to this problem. We have and are running a number of studies in Asian countries to update or define the incidence of red cell antigens and the incidence and specificity of alloantibodies. PM me if you would like to know more or are interested in participating in a study I am planning.

Dear TimOz,

Very thanks for sharing the useful information. Your description was really enlightening!

We have decided to implement the imported panel with 'retrospect' use of "in-house" panel, in case something comes positive while screen is non-commital.

We would love to be a part of your study group. Kindly provide your mail I.D.

Mine is bloodbank.abmh@adityabirla.com

Sorry for joining the discussion a bit late...but I could not resist from joining this interesting thread. The companies which market their products have various reasons to promote or not to promote Indian/Asian/Caucasian cell Panels.

The best way to go forward is to have a metacentric study in India. At least 12-15 Institutes from different parts of the country should participate and data should be looked retrospectively and prospectively.

Some Blood Banks have been doing antibody screening using these 'foreign' made cell panels for more than 2 years. I am sure they have vital data. Someone form the fraternity has collect all this....and then analyse. From the literature that I have gone through, even the Americans did it this way before they adoped T&S policy.

Also, we have to look at the antigens present in these cell panels, the homozygosity of particular antgens. Its possible that some centres use 2 cells for screenin (for cost cutting) and thus miss certain antigens in homozygous dose, thus AB screen is negative and XM is incompatible. The storage of cell..due to heat or humidity.... might destroy certain antigens.

The problem is, we only look for what we want to find. So we are only looking for Mi(a) and Diego...where as there might be something else as well.

Hi Bloodbanker84,

I agree in general with your sentiments regarding a multicentre trial. There is so little real data from the Indian sub-continent. Especially when the phenotypes found there and therefore the antibodies seem to be very interesting. The published data (collated mainly in Mourant Ed 2) also shows that the many different ethnic populations across the country show a wide variety of antigen incidences. Some of this work was done by Roy Simmons from Australia who is an immunohaematology hero but little known (and on a historical note, I still work in the original lab he did his work in in the 40's and 50's.

The other comment I would make on your post is to agree with you about existing data but also to point out that current data tends to only have records for antibodies of the "usual suspects" - i.e. antibodies detected by screening cells to detect antibodies prevalent in Caucasian and in some cases Negroid populations. In many Asian populations, the real antibodies that are of interest are the ones that cause clinical effect and are higher incidence in Asian populations. Many of these are simply not detected by currently available panels and will therefore be lacking from current data. That is why much of the published data in Asia looks fine and was fine at the time but has holes in it. The HK study by Lin et al was groundbreaking at the time ans really shows what is missed. If you wanted to be critical, they were using what was state of the art at the time but our knowlege now shows us that screening with a so called Mia pos cell of single specificity may not have picked all of the vMNS antibodies in the population. If you wanted to make this study a step better I think that you would attempt to further define the specificity of the antibodies, test the thermal amplitude and class and try to link some cases to clinical effects (if possible). The clinical effects of Dia are quite well defined but it seems vMNS behave like other MNS antibodies in that many of them appear to be non-red cell immune, RT only reactive and likely not clinically relevant - but who really has proven this.

In a place like India (and other places like the Philippines) it is ideal to keep every sample that reacts in an antibody screen and especially those that you find in a crossmatch and try to find out what they are. Only then will you define the incidence of antibodies like Dia and vMNS but also find if there are other uncharacterised antibodies in the population. I think we agree that it is very interesting what these "unidentified antibodies" really are and if they matter.

Hi Tim,

You had mentioned that you are planning a study in Asia. I was wondering if my blood bank can be a part of this study.

You can write to me on my email id bloodbanker84@gmail.com

It is time that we start the process of developing the 'perfect screening panel' for India.

Thanks and best regards,

BB84

Apropos of not a lot, we've just found one of our regular donors (who has given 42 times before) to be Vw+ (give or take, about 0.057% of our population) because she was incompatible with a patient.

We're just testing her for Mi(a) now.

The anti-Mia has gone off.

Sorry to come back in so late. A couple of points I need to make though.

To transfusiongpi, yes, I am connected to DiaMed, and, if my sources are correct, we actually met in Pune where I gave a talk about the importance of antibody screening as well as crossmatching by Coombs. And I agree 100% that until there are properly adapted screening cells transferring over to a type and screen is unwise. As I said then, it is important to try and characterise all samples that give a positive crossmatch where the antibody screen is negative. However the biggest difficulty here is that it is impossible to ship samples out of India to get them to a reference lab that would have the neccessary resources to be able to identify the antibody in question. Once identified, if storage conditions exist the same antibody can be used to screen other donors/patients. Believe me, every time DiaMed people came over from India to Switzerland, they alsways tried to smuggle samples over - but usually by the time they arrived they were completely haemolysed and cooked. (If there are any customs officers reading this, please forget I've written this!)

As for the Asia panel, this was a 3-cell panel with all the 'important' antigens plus an 'Mia+' cell. Unfortunately it's not so easy to get hold of the cell, and it's temporarily unavailable.

But I would also like to query the statistics that have been put forward. As I understand it, these figures were presented at the Indian ISBT meeting earlier this year. Is that correct? Please correct me if I'm wrong. If so, there were 1200 samples tested, in which something like 8 antibodies were detected by both antibody screen and crossmatch and only 1 antibody was found in the crossmatch only. I don't know if the antibody was identified.

Hi Anna,

I am not sure if you are asking about the published statistics I quoted earlier. If so, please let me know.

If the statistics you have quoted above are correct than I think it is worth looking at what this may mean. IF vMNS antibodies are present in the Indian population (and this has not been proven as far as I can tell) and IF the incidence of vMNS antigens are present at say 3% (not proven in India but tested in Indian ethnic populations in Malaysia). An antibody screen lacking the ability to detect vMNS antibodies will detect no vMNS antibodies. Assuming a 1 unit crossmatch every time, a crossmatch will statistically only detect 3% of vMNS antibodies. A vMNS antibody will only be seen when by chance a vMNS antigen positive unit is selected for crossmatch.

So, the true incidence of alloantibodies in the population may be 30 times that reported in such a study. I am not criticising such a study at all, just pointing out that such a study has gaps and these gaps may be vaery large. The study by Lee et al in 2003 showed exactly this. Half of all alloantibodies in antenatal females in Hong Kong would be missed entirely by a study using Caucasian based screening cells and crossmatch.

Dear Tim - my mail was not directed to you. I don't think that WE'VE ever met! It was destined for transfusiongpi. sorry about the confusion. And I agree about the need to carry out a large scale study which would take into account the different regions of India rather than just taking India as a whole. Likewise for Africa.

Anna

Though I am a long way from India, I find this discussion very interesting. Becaue we are a university town, we have a fairly large population of Indians as well as Chinese and other Asians in our community.

I have often wondered if we miss antibodies when we test them. But thinking about it more, I wondered if they are exposed primarily to blood from European Caucasians, they are not likely to be exposed to these antigens and not likely to make these antibodies...??

Linda Frederick

That is probably true Linda, but they would be exposed to the c antigen, which is comparatively rare in the Asian population, and is, at the same time, the second most immunogenic antigen of the major blood groups after D. I wonder how many make anti-c, especially females of child-bearing potential?