A long time ago now, I was working in a very large London teaching hospital, when we received a patient who was a group B police officer (this was about the time when the IRA were active in London, but this case, as far as is known, ha nothing whatsoever to do with them). He had received multiple stab wounds.
We soon went through our stock of group B, and eventually got through out stock of group O. Although we had ordered more stock to be delivered by "Blues and twos", at this stage we had a choice. We either transfused him with group A, or we let him die. The doctors in charge decided to give him group A.
He survived, and when the emergency order of stocks has arrived, we switched back to group B, and then group O. Yes, his renal function was shot to pieces for a while, but, to be honest, that was probably the least of his worries at the time. I'm not saying that this would work every time, because it won't, but you can treat a haemolytic transfusion reaction, even an acute haemolytic transfusion reaction; death is difficult to treat.
As Prof Brian McClelland MB ChB ND Linden FRCP(E) FRCPath (former Director of the Scottish National Blood Transfusion Service) once wrote in Thomas D, Thompson J, Ridler B. A Manual for Blood Conservation. 1st edition. 2005. tfm Publishing Ltd, "Transfusion has risks, but bleeding to death is fatal."!
To my own shame, I once did a book review of this for the BBTS, and misquoted the title as, "A Manual for Blood Conversation."! The embarrassment!