I was thinking this phenomenon could be a part of transfusion reaction caused by the ABO incompatible red blood cell transfusion. However, in order for us to report a valid DAT result, the saline control should be negative. We could try to use the warm saline to wash the cells, then the DAT including the saline control might become negative. I doubt the negative result would represent the true situation of the transfusion reaction. How do you think?

I have, unfortunately, seen two ABO transfusion reactions, both of which were fatal.  In both cases, the DAT was negative, BUT, when the blood samples were put under the microscope, the reason was only too evident.  There were hardly any red cells in the samples, presumably because the complement system had haemolysed both the transfused red cells, but also the "innocent bystander" autologous red cells.

That having been said, your explanation is perfectly logical.

An interesting case/issue. I agree with your premise: doing anything to make the saline control nonreactive (warm-washing, CDP, acid-elution) will probably make the test with anti-IgG negative, too (assuming the positive saline control is due to IgM-coating of the patient cells, causing direct agglutination).

Perhaps a DAT with an anti-IgM reagent could be arranged ? Did you try a DAT with polyspecific antiglobulin reagent or a monospecific anti-Complement reagent ? But even results from these could be "invalidated" by reactions in the so-called negative control.

I would wash the red cells in saline and test the DAT (I have occasionally found stronger reactions).
You can also try washing the red cells using cold Elu-Wash as that can help bind any weak antibodies to the red cells.

In this case I would do an eluate including A1, A2 and B cells irrespective of the DAT results.

Lastly I assume you can call the transfusion reaction irrespective of DAT results if hemolysis is evident in the post sample?

The only hemolytic transfusion reaction I worked-up was clear cut and involved uncrossmatched blood given to a patient with history of an anti-Jk(a), against the BB tech’s advice. As Malcom stated above not many red cells were left, including the patient’s as the hemoglobin went from a 6 to a 3! I could not tell where the plasma ended and red cells begun. 

50 minutes ago, Ensis01 said:

In this case I would do an eluate including A1, A2 and B cells irrespective of the DAT results.

Excellent point, Ensis01.

Thank you all for responding to this topic. Fortunately, the patient survived probably due to bleeding at that moment. Both poly- and mono-specific DATs were done and they (i.e. poly DAT, IgG, complement and saline control) were all weakly positive. An eluate was tested against A1, A2 and B cells and reacted with A1 and B cells, not A2 cells. The reaction strength with A1 cells was stronger than B cells. The possible explanation for eluate reacting with B cells is that the anti-A,B in this O patient's plasma coated on the A donor cells got eluted and reacted with group B reagent cells.