What are your rules for ruling out?

[mwilde]

I actually have a wonderful rules page I got from a retired hematopathologist. But, I don't know who to post that here. If anyone would like a copy, email me: lthedford@petersonrmc.com

I would love to see the page you have. Please email it to me at: mewilde1@earthlink.net and mwilde@lakelandregional.org.

Thanks a bunch!

Margaret Wilde

[margaretcox]

We of course use more than 1 (say, 3) negative and 3 positive cells to rule IN an antibody. That's just statistics, to ensure a decent probability that the antibody is what you say it is. But to rule out: only 1 homozygous. You rule out plenty of antibodies with 1 homozygous cell every time you report a negative antibody screen.

[janet]

For those using 2 homozygous when possible .... do you use that rule for the "clinically insignificant" M (Lewis and P for that matter too)

Our rule is 2 M, one for Lewis and one P but one of my technologists is asking me why .... we just give IAT compatible units for those antibodies so why bother ruling them out when in the end the crossmatch will tell all.

I agree with here but am looking for other's thoughts for or against :0)

[Malcolm Needs ☆]

No, in fact, in most cases, we don't rule them out at all.

For example, if we perform a panel, and we find an anti-K present, and the K+ red cells happen to be the only Lu(a+) red cell on the panel, we would not even bother to run a cell that is K-, Lu(a+), simply because anti-Lua is not clinically significant.

[jcdayaz]

Our "rules" are 1 homozygous or 2 heterozygous....with the exception of the Kidd, Duffy, and MNS family of antibodies. Those need homozygous cells to rule out. I have seen MULTIPLE MNS system antibodies react only on certain homozygous cells....even some homozygous cells might not react....can be challenging....

[Malcolm Needs ☆]

Our "rules" are 1 homozygous or 2 heterozygous....with the exception of the Kidd, Duffy, and MNS family of antibodies. Those need homozygous cells to rule out. I have seen MULTIPLE MNS system antibodies react only on certain homozygous cells....even some homozygous cells might not react....can be challenging....

Ah, but don't forget there are some 46 antigens within the MNS Blood Group System, many of which are the result of genetic "cross-over" (either Lepore or anti-Lepore types) and some of these are quite common in certain areas of the world (see many of TimOz's posts). It is probable, therefore, that certain M+N- and/or M-N+ apparent homozygosity will be, in fact, and for example, M+Lepore type+.

:confused:

[jcdayaz]

We use an antibody identification software to choose our select cells for ruling out/proving...whatever. No more need to spend potentially hours flipping through a notebook with panel sheets looking for the cells you need!!

The software is automatically updated with each new lot/panel of reagents you get. You simply enter what antigens you want positive or negative, and if you want them homozygous or heterozygous and presto! the program tells you which panel to use, which cell within that panel--everything to make your search MUCH easier.

Check it out.....www.AntibodyCheck.com

[Joanne P. Scannell]

Interesting and SCARY replies! I see serious issues here:

1. 'Rule of Thumb' here in the USA, 3 positives or 3 negatives to put you on the cautious side of the statistical probability that the identification is correct. If two are working for you 'across the pond', that's wonderful.

This does not mean you need 6 cells. It means if you have an antibody to a high incidence antigen, you need 3 negatives to make the ID ... likewise in reverse for the lower incidence antigens.

e.g. We don't call it an Anti-E unless we have 3 positive (and you know there are plenty of negatives around). We don't call it an Anti-k unless we have 3 negatives (again, with plenty of positives around).

Besides, if you insisted on 3 pos and 3 neg to rule in/out ... ummm, how do you get around calling an Antibody Screen negative?

2. Zygocity: We have met many antibodies (eg. Anti-D, -E, Jka) on many occasions that react with ONLY homozygous cells. You CANNOT rule out with heterozygous cells as a general rule! There have been articles and lectures over the years about the dangers of doing this. However, there are a few that can 'sneak' by, eg. Anti-K1, because they generally do not succumb to dosage effects.

Caution: Kkeep in mind that cells that may 'look' homozygous on the panel, eg. Fya+b-, may not be homozygous ... that 'second seat' may be occupied by an alternate antigen, e.g. Fy3.

3. 'One Homozygous Cell is Enough' : When we run an Antibody Screen, we are 'ruling out' using ONE homozygous cell all day, every day and we feel safe saying there are no antibodies and perform immediate spin crossmatches. Why should this change when running a panel of cells, then?

So, no ... we don't rule anything out (except Anti-K1) unless we have a homozygous cell that is not reacting with our 'unknown antibody'. Those that cannot be ruled out using a homozygous cell are listed as 'possble Anti- ___' until presence is either proven or dismissed. And yes, we 'honor' them as if they were there until proven otherwise ... with educated prudence, of course!

[jcdayaz]

No, in fact, in most cases, we don't rule them out at all.

For example, if we perform a panel, and we find an anti-K present, and the K+ red cells happen to be the only Lu(a+) red cell on the panel, we would not even bother to run a cell that is K-, Lu(a+), simply because anti-Lua is not clinically significant.

It was presented at an AABB conference several years ago (sorry, can't remember precisely which year) that you can rule out antibodies if you have reactions (any strength) and your screen &/or panel are totally negative for that particular antigen. Typically Antibodies that are usually easily excluded following these guidelines are Jsa, Kpa, Lua, and probably more I am forgettting because I don't have a panel sheet in front of me. We follow these AABB guidelines.

[Malcolm Needs ☆]

It was presented at an AABB conference several years ago (sorry, can't remember precisely which year) that you can rule out antibodies if you have reactions (any strength) and your screen &/or panel are totally negative for that particular antigen. Typically Antibodies that are usually easily excluded following these guidelines are Jsa, Kpa, Lua, and probably more I am forgettting because I don't have a panel sheet in front of me. We follow these AABB guidelines.

If you do remember, I'd love to see an abstract of the presentation.

I'd be a little worried about the anti-Jsa, but would agree that the others tend to be clinically benign, howver, you can't exclude them, just ignore them..

[jamato]

We almsot got burned with this rule. It involved an anti-S. The tech did not see the pattern with the homozygous S. She called it an anti-M. It was picked up upon review before the patient received any units.:fear:

[jcdayaz]

We almsot got burned with this rule. It involved an anti-S. The tech did not see the pattern with the homozygous S. She called it an anti-M. It was picked up upon review before the patient received any units.:fear:

oooppps

Edited November 18, 2009 by jcdayaz

[jcdayaz]

ooopppss

Edited November 18, 2009 by jcdayaz

[SMILLER]

One **** and two hetero, if available. On non-dosage antigens, we would allow just 3 heteros. Obviously an AG can be typed positive on the patient we would rule those out as well.

[Mabel Adams]

But every day on your negative antibody screens you rule out antibodies with usually a single homozygous cell and for K a single hetero cell. (As you may notice the "censors" didn't like your abbreviation for homozygous. ) Does anyone have any logic for why some of us want to use so many more rule-out cells when the patient is known to have an antibody? Is it because they are known "responders"?

[lab217]

I am finding, since switching from tube testing to solid phase testing, that these rules are becoming more and more difficult to follow. You are bound to what the manufacturer decides is helpful in the panel. We are finding that solid phase testing is so much more sensitive, could these rules be changing?

In these type of testing particularly, we are finding that in really difficult cases, the solid phase crossmatch is the most helpful piece of information.

[jcdayaz]

Um...

My lab strictly uses 3 homos... and 2 heteros can replace one **** if absolutely necessary...

What's your take on my lab's methods? =S

Well, hmm. I think the term "strictly" just might not apply here. Are you saying if you have something like an Anti-k your policy dictates that you have to find 3 homozygous cells to prove it? We keep 6 months of expired panels to use for select cell purposes. I'm not sure we would be able to find 3 homozygous k cells in our six month inventory.

[Mabel Adams]

There is a difference between what you require to rule out a specificity vs what you require to ID a specificity. The old 3 neg & 3 pos rule is for IDing and sometimes is confused as being necessary for ruling out. Obviously, we rule out every day on neg screens with 1 hetero K cell and mostly 1 homozygous cell for most other specificities.

Statistically, if you have 3 pos cells reacting and 3 neg cells not reacting (zygosity should not matter) the antibody is likely to be that specificity because it is very unlikely that you would get that same pattern randomly. You can use 2 pos and 5 neg or 5 pos and 2 neg to reach the same statistical degree of likelihood. Again this it to ID, not to rule out.

[Gnapplec]

Our reference lab just recently changed to a more strict rule of 3 cells (two homozygous will no longer suffice).

In certain cases, where there are multiple antibodies, the rules are bent a little. But for now = as many homozygous cells as physically possible... 3 to rule in/out is the new rule.

[webersl]

...(As you may notice the "censors" didn't like your abbreviation for homozygous. ) ....

I am so beyond annoyed at that. They cannot hijac a greek prefix in common scientific use and decide it is now unprintable! I'd like to issue a formal complaint.

Now back to the point. We may need to rethink our rule-out rules in the light of solid phase and gel. At our hospital, we start with gel. If an antibody is showing dosage in gel - giving a 1+ on poof cells (+/-),but negative reactions on straight cells (+/+), it is nearly certain that we would not even see this antibody at all in tube. If you had started in tube, you might have ruled it out 3 times on a poof cell, but the fact is, it was actually there, but too weak to see. So, my point is, there is a continuum. We give out our results as "positive" or "negative" for a certain antibody, but the fact is, what we are truley observing and should be reporting is "below the limit of detection" or "above the limit of detection". Different methods (and even different cells) have different limits of detection. So, I'm saying that if various tube methods are the gold standard for the limit of detection, perhaps those using more sensitive methods could have different rule-out/rule-in rules than those using tube. Thoughts? (not about orientation, just the science part, please)

Edited April 13, 2012 by webersl
spelling

[SMILLER]

Great, now they will start censoring "poof".

Next thing, we will be noting on unit crossmatch tags: "Compatible-As far As We Can Tell"

[webersl]

...

Next thing, we will be noting on unit crossmatch tags: "Compatible-As far As We Can Tell"

I LOVE this!

[Malcolm Needs ☆]

Well, in a way, it's true!!!!!!!!!!!

[lacs]

1 homozygous

1 homozygous and 1heterozygous

2 homozygous

anything goes

2 homos to rule in and out.

[kirkaw]

I had a case today, where a patient had a known anti-M identified previously at another hospital. Our tech had already set up a panel before we found out this information. The panel showed various reactivity...not something that I attributed to dosage. She had ruled out anti-S using 3 heterozygous cells. I asked that we pull an S homozygous, M- cell for rule out. Is this overkill?

Thanks!