How exact do we use fisher?

[Rh-fan]

Beside the discussion on the need to use the fisher-exact calculation, I have a question on how to use it. When I have to use it (because of regulations), I want to use it right.

For one antibody it is clear and easy, but with 2 or more it gets complicated.

when not all antibodies are excluded (let say a anti E beside an suspected anti Fyb), do you use the Fyb+ E+ cells as positives for the Fyb or not?

I would say first exclude everything on the panel and then calculate the fisher-exact. When it is not possible then you should not calculate with these cells as positive, just leave them out of the formula.

Is there anybody who has an opinion on this.

[Malcolm Needs ☆]

We would use an E-, Fy(b+) and an E+, Fy(b-) as positives.

[John Eggington]

Also, exclusion don't 'stand alone', so, in this example, if the patient was E+ and/ or an enzyme panel was 'negative' would you still 'exclude' anti-E?

[Malcolm Needs ☆]

Not necessarily so, I'm afraid John. There have been examples of anti-E that react by IAT, but not with enzyme-treated red cells, e.g.

MacRate EB, Carrington T, De Lord C, Needs M, Win N. Anti-E detectable by IAG and not by enzyme? How confusing! Transfusion Medicine 2001; 11 (suppl 1): 30.

[John Eggington]

I was just making a general point, that this particular example highlighted - that there are other things to think about. I know some antibodies, that you'd expect to be detected by 'enzymes' (the cited anti-E, for example), may not react as you'd like, but they're not common enough to make us doubt our 'enzyme' panels, I hope (which is why they get published, I guess - who wants to read about ant-E that was detected by 'enzymes'!). Hooves and zebras, as is oft quoted here...

[Malcolm Needs ☆]

I was just making a general point, that this particular example highlighted - that there are other things to think about. I know some antibodies, that you'd expect to be detected by 'enzymes' (the cited anti-E, for example), may not react as you'd like, but they're not common enough to make us doubt our 'enzyme' panels, I hope (which is why they get published, I guess - who wants to read about ant-E that was detected by 'enzymes'!). Hooves and zebras, as is oft quoted here...

True, very true!