Confused about dosage

[kitty 10]

My lab has decided that all generalist are to perform antibody panels and ID's. Keep in mind that I have been a generalist for 25 plus years and can perform the basic antibody screen but have never had to ID the antibody, we would send those to a reference lab.  As with time so goes the memory, so I'm asking if  someone would  please explain "dosage" on a antibody panel and how to rule them out.. I just want to make sure I have a clear understanding of the rule out procedure. Thanks for your help...

[David Saikin 1,353]

Dosage addresses the expression of ag on the red cell.and its reactivity with antibody.  Homozygous intimates a single expression of the gene.  Let's say we/re talkijng about the K ag.  KK is homozygous for K, Kk is heterozygous for both K and k , kk ia homozygous for k.  Dosage occurs when the antibody reacts less strong when the gene products are heterozygous, i.e, the homozygous expression will display stronger reactions.  

The systems which express dosage are the Rh, MNSs, Kidd, and Duffy.  The texts tell you the Kell system ags do not express dosage but I have found the reality is that they do.

When ruling out antibodies it is generally considered good practice to not rule out based on a negative result with a heterozygous cell. There are modalities of testing which enable the use of heterozygous cells for rule outs:  enzyme pretreatment (not for Duffy or MNSs);  I've also considered PeG to be valid for this.  

Hope this helps

 

[R1R2 292]

There are some nifty antibody id programs out there, like Antibody Check, that can aid in the identification of abs.   I think this would be a great tool for generalists. 

 

[SMILLER 860]

Just to add a bit to what David has already explained.  I tend to think of dosage as relating to the amount of antigens present on the RBCs that you are using to ID the patient's antibody, and if the reagent RBC has lots of antigens of the type in question, then the reaction will be stronger.  This is really important for a patient whose antibodies are just developing--you want to use a reagent RBC with the strongest expression possible, and these are the homozygous cells.

For example, at our hospital, we use the 3 by 3 method for antibody ID (for each type of significant antibody, if the antibody is present, we want to rule in with 3 positive RBCs, and to rule out all the other antibodies, we want to have 3 negative reactions for those.)  So for antigens that "show dosage", we want at least one of those three rule out RBCs to be homozygous.

Scott

[Malcolm Needs ★ 4,689]

44 minutes ago, SMILLER said:

Just to add a bit to what David has already explained.  I tend to think of dosage as relating to the amount of antigens present on the RBCs that you are using to ID the patient's antibody, and if the reagent RBC has lots of antigens of the type in question, then the reaction will be stronger.  This is really important for a patient whose antibodies are just developing--you want to use a reagent RBC with the strongest expression possible, and these are the homozygous cells.

And just to add a bit to what Scott said, I will give an example.

There are, on average 14, 000 Kidd antigens per red cell.

This means that for a red cell with the phenotype Jk(a+b-), with a probable genotype of JKA/JKA, all 14, 000 Kidd antigens would be Jk^a.  Similarly,in the case of Jk(a-b+), with a probable genotype of JKB/JKB, all 14, 000 Kidd antigens would be Jk^b.

In the case of a red cell with the phenotype Jk(a+b+), with a probable genotype of JKA/JKB, 7, 000 of the Kidd antigens would be Jk^a, and 7, 000 of the Kidd antigens would be Jk^b.

In the case of a weakly reacting anti-Jk^a, showing dosage, the antibody would react with the Jk(a+b-) red cells, but would not necessarily react (visually) at all with the Jk(a+b+) red cells - or may react with them, but the reactions would be distinctly weaker than the reactions with the Jk(a+b-) red cells.

In the case of a weakly reacting anti-Jk^b showing dosage, the antibody would react with the Jk(a-b+) red cells, but would not necessarily react (visually) at all with the Jk(a+b+) red cells - or may react with them, but the reactions would be distinctly weaker than the reactions with the Jk(a-b+) red cells.

I hope this helps, and doesn't serve to muddy the waters more.

[Marianne 65]

I agree that looking into software programs to build in rules and help would be good for a lab without a dedicated blood banker.  AntigenPlus is another good option

[Texas Lynn 14]

We may start doing panels here also. Where do I get these programs to ID the antibody?

[David Saikin 1,353]

40 minutes ago, Texas Lynn said:

We may start doing panels here also. Where do I get these programs to ID the antibody?

I cannot tell you where to get a program.  As I recall, you will need to enter your antigram into that program.  Some panels may come with downloadable antigrams.  JUst thought you should be aware of this.

[R1R2 292]

48 minutes ago, Texas Lynn said:

We may start doing panels here also. Where do I get these programs to ID the antibody?

I have used and liked this one-

http://www.antibodycheck.com/

 

[kitty 10]

Thanks to all who responded, I think I have a clearer understanding as I go forward, I just need my brain to re-learn all of this... I will have my Lab Director look into some of these programs that you have suggested.  Again thanks

[MaryPDX 39]

RC-Aid is another program. 

http://www.rcaid.net/

 

[AMcCord 1,224]

I also like AntibodyCheck. You receive files by email monthly with the antigrams for your screen and panel cells. Those files are downloaded/saved in files accessible to the software and it does the rest. 

[SMILLER 860]

From the My Two Cents Dept...

I would just point out that it is important that people doing testing understand what and why they are doing what they are doing.  I guess this goes without saying.  I am not a fan of throwing computer AI at problems when staff have trouble understanding what it is they are doing.  I get it that with staff shortages and what not, that generalists have a lot of hats to wear, but a computer algorithm should never be a substitute for appropriate education and regular, effective performance evaluation.

Scott

[AMcCord 1,224]

20 hours ago, SMILLER said:

From the My Two Cents Dept...

I would just point out that it is important that people doing testing understand what and why they are doing what they are doing.  I guess this goes without saying.  I am not a fan of throwing computer AI at problems when staff have trouble understanding what it is they are doing.  I get it that with staff shortages and what not, that generalists have a lot of hats to wear, but a computer algorithm should never be a substitute for appropriate education and regular, effective performance evaluation.

Scott

Absolutely!

[Malcolm Needs ★ 4,689]

On 2017-5-9 at 4:53 PM, SMILLER said:

From the My Two Cents Dept...

I would just point out that it is important that people doing testing understand what and why they are doing what they are doing.  I guess this goes without saying.  I am not a fan of throwing computer AI at problems when staff have trouble understanding what it is they are doing.  I get it that with staff shortages and what not, that generalists have a lot of hats to wear, but a computer algorithm should never be a substitute for appropriate education and regular, effective performance evaluation.

Scott

Hear, hear Scott.

How can an algorithm tell the difference between an anti-D+C and either an anti-C+G or anti-G, anti-hr^B, rather than anti-C+e or anti-hr^S, rather than anti-ce (anti-f), to name but a few?  The answer is that it cannot, and these specificities are much more common than a lot of people think.

[R1R2 292]

On ‎5‎/‎13‎/‎2017 at 2:20 PM, Malcolm Needs said:

Hear, hear Scott.

How can an algorithm tell the difference between an anti-D+C and either an anti-C+G or anti-G, anti-hr^B, rather than anti-C+e or anti-hr^S, rather than anti-ce (anti-f), to name but a few?  The answer is that it cannot, and these specificities are much more common than a lot of people think.

Most programs will not ID the antibody.   They are just a tool for the blood banker to use during ID.     In my experience, most blood bankers, including SBBers could not tell the difference between  an anti-D+C and either an anti-C+G or anti-G, anti-hr^B, rather than anti-C+e or anti-hr^S, rather than anti-ce (anti-f) IMO (sad but true).  

[dot 3]

I like bbguy.org it's a good learning sight.  It helped me learn.  Good luck!

[Malcolm Needs ★ 4,689]

39 minutes ago, dot said:

I like bbguy.org it's a good learning sight.  It helped me learn.  Good luck!

I'm afraid that I have to disagree.  bbguy.org is not good - IT IS BRILLIANT!

[dragonlady97213 84]

We use AntigenPlus, but ONLY for selecting panel cells.  I understand you can enter results into the program and it will let you know what you can and cannot r/o, but that functionality hasn't been validated for use (and I'm glad it hasn't).  IMHO, automation and algorithm's have dumb-downed our profession some, but that's another whole topic and can of worms.  AntigenPlus is a godsend, though, for making selected cell panels since it keeps track of our 9-12 in-date commercial panels and I don't have to rifle through all those antigrams.

[Malcolm Needs ★ 4,689]

7 hours ago, dragonlady97213 said:

IMHO, automation and algorithm's have dumb-downed our profession some, but that's another whole topic and can of worms.

I could not agree more with you.

[diplomatic_scarf 2]

error

Edited May 18 by diplomatic_scarf
error

[Johnv 14]

I have to agree with you too.  The BBGuy was one of my professors during graduate studies and SBB training -  way back when...