SbbPerson

@Neil Blumberg, I wish we had you at all of our facilities to educate our medical staff. Sadly, convincing Hematologists and Oncologists (at least here in America) that it is better to postpone platelet transfusions than give ABO incompatible platelets is, more often than not, rejected, especially in light of the fact that many patients are having to wait because of lack of platelet inventory to begin with. 
What we really need is a push for better transfusion therapy education in medical school. Along with this, continuing education for practitioners needs to become a priority. It is, however, quite difficult to get time with these practitioners. Even when we convince our laboratory medical directors to advocate for these issues, in my experience, clinicians rarely change. 
All that said to say, in the "trenches," the practice will likely continue to prioritize inventory over safety. 

Short answer would be any ABO type if a one time thing, along with a prayer card for no hemolysis or post-transfusion purpura.
 
You could make a case for type A as the anti-B is likely to be lower titer, lower biologic activity than the anti-A in group O platelets (unless low titer) or group B platelets. But this is largely theoretical hand waving.

I should add the good news is that when one starts prioritizing ABO identical platelets over inventory management, one reduces the platelet transfusions needed by perhaps 50%.  So our platelet shortages will disappear in large part if we stick with ABO identical as much as possible.  See attached randomized trial from eons ago :).  ABO identical reduces transfusion reactions as well, HLA and rbc alloimmunization.  Not to mention decreasing bleeding and mortality.
ABO randomized trial UR european j haematology 1993 copy.pdf ABO plt tx revisited cumulative effects.pdf Platelet transfusion worsens ICH Stroke 2020 copy.pdf

Another point.  Since group O whole blood has proven as safe or even safer than typical component therapy (A platelets, A or AB plasma) in massive transfusion of trauma patients, perhaps group O low titer platelets would be safer than group A or B platelets for an AB patient :)?  No one knows, but worth considering.  The big problem is probably giving non-O platelets to O patients. There is evidence this increases bleeding and mortality.  Just like red cells, only O platelets for O recipients is a good practice.  The AB patient may be less of a problem, since giving some small amount of antibody may be less dangerous. A risk of hemolytic reaction of about 1 in 700 or so.  The risk of mortality in transfusing an O patient with A platelets is probably 1 in 5 (see attached).
ABO incompatible platelets intracranial bleeding 2021.pdf ABO plasma incompatible platelets and hemolytic reactions.pdf

"Since AB+ people are considered the "universal recipient" , we give them any type platelets, usually starting with the one with the closest out date. "
I grant you that this is widely shared idea in our field for decades. It is also seriously wrong.  It prioritizes inventory management over patient wellbeing.  Our approach to ABO and platelets is distinctly different from ABO and red cells with no rational basis.  Antibody and complement destroy red cells and platelets equally well.  The only difference is that instead of free hemoglobin being released, it's mediators such as VEGF, IL-6 and other platelet pro-inflammatory, immunomodulatory and pro-thrombotic granule contents are released.   
ABO mismatched platelet transfusions at least double the refractoriness rate in repetitively transfused patients (see attached for references), and actually increase bleeding and mortality. 
The answer to the question is ABO identical is by far most effective and safest.  If you have to give ABO mismatched, there is probably no good answer other than washed/volume depleted O's, A's or B's, where most of the incompatible plasma is removed.  If that's not possible, postponing platelet transfusion until ABO identical is available when feasible, giving half doses of ABO identical if two patients need the one available unit, etc. are also reasonable.
Sadly, ABO mismatched platelets are probably worse than no platelets at all. They provide little or no hemostatic benefit and increased risks of bleeding, organ injury and death for the patient.  If I were the attending physician, I would generally give no platelets if ABO identical or washed O's weren't available in a stable, non-bleeding patient with a count of over 5,000.
The good news is we can improve outcomes by just doing what we do for red cells. Do not transfuse ABO incompatible antigen or antibody. It's bad for red cells, platelets and endothelial cells, all of which have complement and Fc receptors that bind immune complexes, and all of which bear ABO antigens on their surfaces.
Carr ABO mismatched refractoriness copy.pdf ABO story expanded.docx ABO endothelial cell paper.docx NEJMc2034764 copy.pdf NEJMc2034764_appendix copy.pdf

I know Joint Commission doesn't allow boxes on the floor. We had to add shelving with solid bottom shelves to accomplish that. We also are not supposed to store anything under the sinks. We were not cited by CAP for that in the core lab or Blood Bank prior to adding the new shelving. Would be interested to see which checklist item is cited for that, something new?.

Absolutely! It is in our policy in accordance with CAP and AABB standards.

Do you do QC on your expired panels when you use them as selected cells? I was just curious. Thank you

I sincerely hope that you would also contact your blood supplier to let them know that there is a high risk of bacterial infection, as there may well be other blood components from the same donor that may yet still be untransfused to a patient, and which should be immediately put into quarantine.

To my understanding: if you are referring to PAS (PR) platelets, which are being, or have been phased in by blood suppliers; then give any ABO type as 60-70% of plasma has been replaced with crystalloid nutrient media. The pathogen reduction (PR) negates the CMV and irradiated necessity. 
Other BB may have a different policy though platelet availability may give you no choice. 

Since AB+ people are considered the "universal recipient" , we give them any type platelets, usually starting with the one with the closest out date. 

CAP simply requires adequate space, but workload and staff safety are both considerations as a part of that requirement. I agree with pushing the safety angle, as well as instrumentation requirements. 

In this case, the Safety goons may be your very best friends. If the space you describe is really that bad, you could use the safety angle as leverage. Alternatively, modern instrumentation (including the Ortho Vision) often has very specific installation requirements (space/clearance/ventilation, etc)....that may ammunition, too.

To my knowledge, there has never been such a rule, regulation or requirement (the 3 "Rs" that rule the blood bank).  In two blood banks / transfusion services I worked in, one was in a separate room but certainly big enough.  In the other, we were part of the lab but separated from the rest of the lab in the same way chemistry was separate from hematology.  The only department that had it's own room was microbiology.

There is a recommendation, or maybe a standard I can't put my finger on right now, that says that access to the blood bank should be controlled. Meaning people can't just walk in and take blood. Nothing that says it must be a separate room. My entire lab is behind badge access doors, and the blood bank is semi-separated by a counter and some swinging half doors (saloon doors) so that people picking up products would certainly be noticed if they just popped into the actual department. 

You won't be cited for not being in a separate room.  You can be cited if your work area is too small and cramped.  We were cited by the FDA years ago.  I was walking beside the FDA inspector talking to him and one of my staff walked right into me.  
There was also an issue where if one of the blood refrigerator doors was open, there was no way to get from one side of the Blood Bank to the other.
Even so, Administration was not happy at all about having to give us more space.  They did, but grudgingly and they gave us the least amount they could get away with.

There is no hard rule that Blood Bank has to be in a separate room. The whole lab here is in the same room, including the Blood Bank. 

I agree with both Bet'naSBB and jayinsat in that it is probably an antibody directed against a low prevalence antigen.  The problem with identifying the specificity of such an antibody is that there are so many!  To make certain that it is not a "fool's errand", it might be worthwhile trying to get a sample of blood from the putative father, if he is available and/or known.  As the baby is, like the mother, group O, there is a 50% chance that the father will also be group O, in which case it is simple to see if his red cells can be sensitised by a maternal antibody.  If he is not group O, everything is not lost as, as jayinsat suggests, an eluate from the baby's red cells should be clear of all anti-A and/or anti-B.
If the putative father's red cells are compatible by all methods, either there is another explanation for the positive DAT, or he is not the father (or both).

The other thing that springs to mind is that, even if there is an antibody directed against a low prevalence antigen, as you have not identified a specificity using your standard panel, and should the baby develop a clinically significant case of HDN (it is too late for HDF) and require a transfusion, acquiring crossmatch compatible blood, suitable for the baby, should be a simple task.

I did forget to mention that I did do this prior to performing the eluate and testing against the Father's RBC's. I was able to find 8 low frequency antigens (Dia was included) but they all came out to be negative. Unfortunately that was all the low frequency antigens I could find on the panels that we have available here.

I had this very scenario about a year ago and it turned out mom had an anti-Dia. It was not on any of our in lot screening or panel cells. I did as I suggested and ran a select panel against mother's plasma using expired panel cells and identified the Dia. The eluate on the baby was eluted the Dia also. 

If you have enough of the unit left, you could use blood culture vials along with an initial gram stain.

Thanks everyone for the feedback!  ACHC looks like a great alternative to CAP for many reasons.  Might investigate that route.

I have always done joint CAP/AABB inspections and the CAP portion is almost negligible. I don't see a huge pro to dropping CAP if you are already AABB, since the two are very rarely in disagreement, and AABB tends to have stricter guidelines.  

We are currently TJC, AABB and FDA inspected.  They each take a different tack.  AABB requires more auditing and data gathering which we sometimes have difficulty doing with current resources. This site has not been CAP in the 14 years I have been here.  My previous lab switched to TJC from CAP just after I left there.

Malcolm, my very 1st AABB inspection came about 2 months after taking the Blood Bank supervisor job.  After it was over I contacted AABB and told them that I would never let that inspector in my facility again and if they tried to send her I would drop our AABB membership.  YES, the inspection/inspector was really that bad and luckily I never had to carry through with my threats/promises.